↓
Rivya AI Docs

Model API Reference

Check Rivya model IDs, API availability, input modes, parameters, prompt limits, and requirements for signed reference media.

Last reviewed on November 11, 2026

Use this reference to find Rivya model IDs, check API availability, and look up input modes, parameters, prompt limits, and rules for reference media.

The model sections below use the same catalog and public serializer as GET /api/v1/models. The catalog includes models that are listed for discovery but cannot currently be called.

GET /api/v1/models does not require an API key. Build requests with the inputs and modes listed by Rivya; a provider may offer other features.

Generation and Chat requests require an active, unexpired Bearer key for an account allowed to use Public API. The key needs generations:create or chat:create, respectively. Public API must be enabled for the deployment, and the selected model must be available through Public API when the request is checked.

Suno Speech: scenes and supplied scripts

Suno Speech: prompt is the scene in Simple mode (params.customMode="false") or the script in Advanced mode ("true"). Advanced-only settings are tone, variety, vocalGender and backgroundMusic. Omit them in Simple mode; do not send a second script or prompt inside params.

Prompt units and complete layer results

Read prompt_length_unit together with max_prompt_length from GET /api/v1/models. utf8_bytes counts the original text bytes, including spaces and line breaks; utf16_code_units counts JavaScript UTF-16 units. A null limit does not mean unlimited or optional text. Gemini TTS uses the byte unit.

A completed layer-decomposition result can include result.layers alongside the complete result.urls list. Keep each layer’s z_index, URL, size, name, description and bounding box when provided; index zero is the base image. The TypeScript SDK exposes these fields. Upload references for the selected model and pass its original signed metadata; declared dimensions or duration alone do not replace verification.

How To Read This Reference

  • Use the model id as model in POST /api/v1/generations for asynchronous image, video, and audio models.

  • Use chat model IDs in POST /api/v1/chat/completions.

  • Use top-level prompt when the generation model needs prompt text.

  • Use top-level message for Chat API turns.

  • Put fields for the chosen generation model inside params.

  • Check api_input_status and supported_api_inputs before choosing an input method.

  • available_text uses text; available_url_input uses public URLs through the model's declared parameters. These input types do not require a Files API upload.

  • partial_requires_files supports some input without uploading files, while modes that need reference media require Files API uploads.

  • requires_files_api models remain callable when available. Upload compatible reference media through Files API before submitting a generation.

  • Parameter tables show declared keys, types, defaults, ranges, and options. Requirements can depend on the selected mode; check the model's API guide before using reference media.

  • For seedream-5-pro, seedream-5-pro-layer-decomposition, nano-banana-2-lite, qwen-image-3, and qwen-image-3-pro, keep mime_type, width, height, size_bytes, and image_dimensions_token from the original Files API upload response for that model. In params.referenceMediaItems, pass them as mimeType, width, height, sizeBytes, and imageDimensionsToken.

  • Upload again if a required image token is missing or expired. Layer decomposition requires exactly one image whose signed dimensions meet the model's pixel and aspect-ratio limits.

  • Grok Imagine Image 2.0 uses those signed fields for one to five standard image-edit references. With no references, the request uses text-to-image. The aspect_ratio value auto is edit-only; Segment Map and Segment Edit are not available through Public API.

  • Some video modes require signed duration and video metadata. When these are passed in params.referenceMediaItems, map the Files API fields mime_type, duration_seconds, duration_token, video_width, video_height, frames_per_second, video_bitrate_mbps, video_file_size_bytes, and video_metadata_token to mimeType, durationSeconds, durationToken, width, height, framesPerSecond, videoBitrateMbps, sizeBytes, and videoMetadataToken.

  • Use the original upload response for the selected model. Upload again if either required video token is missing or expired; a public video URL alone does not supply signed metadata.

  • Wan 3.0 has three separate seedance_scene choices: text, frames for first/last-frame input, and reference for multimodal references. Do not combine scenes.

  • Wan 3.0 variants are standard and prime; resolutions are 480P, 720P, and 1080P. Its aspect_ratio choices are adaptive, 16:9, 4:3, 1:1, 3:4, and 9:16. audio controls generated audio, and seed sets the seed value.

  • Wan 3.0 duration is an integer from 2 to 30 seconds, or -1 for intelligent duration. Intelligent duration reserves for 30 seconds and cannot be combined with a reference video.

  • Wan 3.0 images need signed MIME type, dimensions, and byte size. Videos also need signed duration and video metadata. Audio duration tokens bind MIME type and byte size, and audio references must accompany an image or video.

  • Wan 3.0 file-to-video and link-to-video inputs are not available. See the Wan 3.0 API guide for its request fields.

  • Model upload limits can differ from generic file-kind defaults. For generation references, include the target model ID when uploading so Files API can check that model's supported kinds, MIME types, sizes, durations, dimensions, frame rates, and bitrates where those checks apply.

  • Check the chosen mode's inputs and parameters after uploading. Files API does not create a generation task or reserve generation credits; generation requests are validated separately.

  • For an available chat model, read chat_capabilities to check non-streaming replies, SSE streaming, image file_id attachments, web search, thoughts, and supported reasoning efforts. Use each capability only when that model supports it.

  • Use the model's API guide for request details. Its model detail page gives product guidance and examples.

Request Pattern

For Z-Image, send a JSON body like this to POST /api/v1/generations:

{
  "model": "z-image",
  "prompt": "A ceramic mug on a light gray background, lit from the side",
  "params": {
    "aspect_ratio": "1:1"
  }
}

For JavaScript and Python requests and idempotency handling, see Create Generation.

When Claude Sonnet 5 is available, use this body with POST /api/v1/chat/completions:

{
  "model": "claude-sonnet-5-chat",
  "message": "Write a plan with three steps for launching a new product photo campaign"
}

For Chat request fields and streaming examples, see Chat API.

Dynamic Model Reference

Use the API guide link for a model's request details and the model detail link for workspace guidance and examples. Request bodies shown here still need valid credentials and any inputs required by the selected mode.

Use this reference with the debugger

Read the quickstart and check the model's fields here. Then try a mock request in the debugger, or sign in for a live request. Live requests can use credits.

GPT Image 2.6 Flare

gpt-image-2-6-flareThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Image · Flare focuses on quick image iteration: turn a written brief into a draft, then refine it with image input. New requests are paused on Rivya.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

GPT Image 2.6 Sunburst

gpt-image-2-6-sunburstThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Image · Sunburst focuses on detailed image composition and editing from text and image input. New requests are paused on Rivya.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

GPT Image 2.6 Economy

gpt-image-2-6-economyAvailable via APIText or URL; Files API for reference media

Image · Create or edit one 1K PNG image. Reference images are optional. Animated GIF output is unavailable on Rivya.

Credit pricing

Credit rate: 5 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

20000 characters

Reference media

image · 16 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "gpt-image-2-6-economy",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspect_ratio": "auto",
    "background": "auto",
    "output_format": "png"
  }
}

Reference rule

Reference images: Add up to 16 JPEG, PNG, or WebP images, up to 30 MB each.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNoauto-auto, 1:1, 16:9, 9:16, 3:2, 2:3, 21:9, 4:3
backgroundselectNoauto-auto, transparent, opaque
output_formatselectNopng-png

GPT Image 2.5 Flare

gpt-image-2-5-flareAvailable via APIText or URL; Files API for reference media

Image · GPT Image 2.5 Flare supports image creation from a prompt and edits with image references, plus aspect ratio, resolution, and background settings.

Credit pricing

Credit rate: 3–16 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

20000 characters

Reference media

image · 16 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "gpt-image-2-5-flare",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspect_ratio": "auto",
    "resolution": "1K",
    "background": "auto"
  }
}

Reference rule

Reference images: Choose JPEG, PNG, or WebP images, up to 16 in a task and 10 MiB per upload.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNoauto-auto, 1:1, 3:2, 2:3, 4:3, 3:4, 16:9, 9:16, 21:9, 27:16, 16:27, 9:8, 8:9
resolutionselectNo1K-1K, 2K, 4K
backgroundselectNoauto-auto, transparent, opaque

GPT Image 2.5 Sunburst

gpt-image-2-5-sunburstAvailable via APIText or URL; Files API for reference media

Image · GPT Image 2.5 Sunburst generates images from a prompt, with optional reference images and separate aspect ratio, resolution, and background settings.

Credit pricing

Credit rate: 3–16 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

20000 characters

Reference media

image · 16 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "gpt-image-2-5-sunburst",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspect_ratio": "auto",
    "resolution": "1K",
    "background": "auto"
  }
}

Reference rule

Reference images: For an editing request, add up to 16 JPEG, PNG, or WebP images. Each file is limited to 10 MiB by the server upload policy.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNoauto-auto, 1:1, 3:2, 2:3, 4:3, 3:4, 16:9, 9:16, 21:9, 27:16, 16:27, 9:8, 8:9
resolutionselectNo1K-1K, 2K, 4K
backgroundselectNoauto-auto, transparent, opaque

GPT Image 2

gpt-image-2Available via APIText or URL; Files API for reference media

Image · GPT Image 2 on Rivya supports text-to-image generation and reference-image edits, with aspect ratio and resolution controls.

Credit pricing

Credit rate: 3–8 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

20000 characters

Reference media

image · 16 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "gpt-image-2",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspect_ratio": "1:1",
    "resolution": "1K"
  }
}

Reference rule

Reference images: Add up to 16 JPEG, PNG, or WebP reference images for an edit. The server upload limit is 10 MiB per file.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNo1:1-auto, 1:1, 9:16, 16:9, 4:3, 3:4
resolutionselectNo1K-1K, 2K, 4K

GPT Image 1.5

gpt-image-1-5Available via APIText or URL; Files API for reference media

Image · Create images from a prompt with optional reference images using GPT Image 1.5. Choose Medium or High quality and an aspect ratio for product, campaign or scene drafts.

Credit pricing

Credit rate: 4–22 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

-

Reference media

image · 16 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "gpt-image-1-5",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "quality": "medium",
    "aspect_ratio": "1:1"
  }
}

Reference rule

Reference image: Attach up to 16 JPEG, PNG or WebP images, each no larger than 10 MiB. Describe the role of each reference in your prompt.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
qualityselectNomedium-medium, high
aspect_ratioselectNo1:1-1:1, 2:3, 3:2

Z-Image

z-imageAvailable via APIText input ready

Image · Alibaba's Z-Image generates images from text. Choose a square, portrait, or landscape aspect ratio for concept drafts, illustrations, and social graphics.

Credit pricing

Credit rate: 0.8 per image The final cost is rounded up to whole credits.

Modes

text-to-image

API input

Text input ready

Inputs

text

Prompt limit

1000 characters

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "z-image",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspect_ratio": "1:1"
  }
}

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNo1:1-1:1, 4:3, 3:4, 16:9, 9:16

Nano Banana 2.1

nano-banana-2-1Available via APIText or URL; Files API for reference media

Image · Nano Banana 2.1 — Create one image from a prompt or reference images.

Credit pricing

Credit rate: 3–8 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

20000 characters

Reference media

image · 10 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "nano-banana-2-1",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspect_ratio": "auto",
    "resolution": "1K",
    "output_format": "jpg"
  }
}

Reference rule

Reference image: Up to 10 JPEG, PNG, or WebP reference images, each no larger than 30 MB.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNoauto-1:1, 2:3, 3:2, 1:4, 4:1, 3:4, 4:3, 4:5, 5:4, 1:8, 8:1, 9:16, 16:9, 21:9, auto
resolutionselectNo1K-1K, 2K, 4K
output_formatselectNojpg-png, jpg

Nano Banana 2

nano-banana-2Available via APIText or URL; Files API for reference media

Image · Nano Banana 2 is a Google image model offered through Rivya for product images, brand visuals, and campaign drafts from text or reference images. Choose the output tier, framing, and format, then compare the result with your brief.

Credit pricing

Credit rate: 5–12 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

20000 characters

Reference media

image · 10 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "nano-banana-2",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspect_ratio": "auto",
    "resolution": "1K",
    "output_format": "jpg"
  }
}

Reference rule

Reference image: Upload up to 10 JPEG, PNG, or WebP images, up to 30 MiB per file (binary megabytes). Include only what the request needs and explain each image's role in the prompt.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNoauto-auto, 1:1, 16:9, 9:16, 3:2, 2:3, 4:3, 3:4, 5:4, 4:5, 21:9, 4:1, 1:4, 8:1, 1:8
resolutionselectNo1K-1K, 2K, 4K
output_formatselectNojpg-jpg, png

Nano Banana 2 Lite

nano-banana-2-liteAvailable via APIText or URL; Files API for reference media

Image · Nano Banana 2 Lite is Rivya's lower-credit Nano Banana option for quick product concepts, social image variants, and reference-led drafts, with a fixed 3-credit generation price and 15 aspect-ratio choices.

Credit pricing

Credit rate: 3 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

20000 characters

Reference media

image · 10 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "nano-banana-2-lite",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspect_ratio": "auto"
  }
}

Reference rule

Reference images: Upload up to 10 JPEG, PNG, or WebP images, no more than 30 MB each.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNoauto-auto, 1:1, 1:4, 1:8, 2:3, 3:2, 3:4, 4:1, 4:3, 4:5, 5:4, 8:1, 9:16, 16:9, 21:9

Nano Banana Pro

nano-banana-proAvailable via APIText or URL; Files API for reference media

Image · Create brand concepts, product stills, and campaign images with Nano Banana Pro. Write a prompt, add optional reference images, then choose resolution, framing, and PNG or JPG output.

Credit pricing

Credit rate: 8–14 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

10000 characters

Reference media

image · 8 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "nano-banana-pro",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspect_ratio": "1:1",
    "resolution": "1K",
    "output_format": "png"
  }
}

Reference rule

Reference image: Add up to 8 JPEG, PNG, or WebP images, each up to 30 MiB. Sign in to upload files.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNo1:1-1:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 21:9, auto
resolutionselectNo1K-1K, 2K, 4K
output_formatselectNopng-png, jpg

Nano Banana

nano-bananaAvailable via APIText or URL; Files API for reference media

Image · A Google image model for text-to-image and image-to-image drafts, with aspect-ratio and PNG/JPEG controls for portraits, product compositions, and wide layouts.

Credit pricing

Credit rate: 3 per request The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image · 10 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "nano-banana",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "imageSize": "1:1",
    "output_format": "png"
  }
}

Reference rule

Reference image: Sign in to upload up to 10 JPEG, PNG, or WebP image references, each no larger than 10 MiB. Use images you have permission to submit.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
imageSizeselectNo1:1-1:1, 9:16, 16:9, 3:4, 4:3, 3:2, 2:3, 5:4, 4:5, 21:9, auto
output_formatselectNopng-png, jpeg

4o Image

4o-imageAvailable via APIText input ready

Image · 4o Image is an OpenAI image model offered through Rivya for concept images, product and brand drafts, and campaign layouts from a text brief. Choose an aspect ratio, describe the composition, and compare the finished image with your brief.

Credit pricing

Credit rate: 3 per request The final cost is rounded up to whole credits.

Modes

text-to-image

API input

Text input ready

Inputs

text

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "4o-image",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "size": "1:1"
  }
}

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Parameters

keyTypeRequiredDefaultRangeOptions
sizeselectNo1:1-1:1, 3:2, 2:3

Seedream 5.5 Text to Image

seedream-5-5-text-to-imageAvailable via APIText input ready

Image · Generate one image from a text prompt with Seedream 5.5. Choose 1K, 1.5K or 2K output.

Credit pricing

The credit cost depends on your inputs and settings. Check the estimate before you start. The final cost is rounded up to whole credits.

Modes

text-to-image

API input

Text input ready

Inputs

text

Prompt limit

5000 characters

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "seedream-5-5-text-to-image",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspect_ratio": "1:1",
    "size": "2K",
    "output_format": "png",
    "background": "auto"
  }
}

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNo1:1-1:1, 4:3, 3:4, 16:9, 9:16, 2:3, 3:2, 21:9
sizeselectNo2K-1K, 1.5K, 2K
output_formatselectNopng-png, jpeg
backgroundselectNoauto-auto, transparent, opaque

Seedream 5.5 Image to Image

seedream-5-5-image-to-imageAvailable via APIFiles API needed

Image · Edit an image with Seedream 5.5 using 1–10 references and a text instruction.

Credit pricing

The credit cost depends on your inputs and settings. Check the estimate before you start. The final cost is rounded up to whole credits.

Modes

image-to-image

API input

Files API needed

Inputs

file

Prompt limit

5000 characters

Reference media

image · 10 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Reference files: Upload JPEG, PNG or WebP images up to 10 MB each and write 5–5000 characters. Choose 1K, 1.5K or 2K and PNG or JPEG. Auto aspect ratio is available; transparency requires PNG.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNoauto-1:1, 4:3, 3:4, 16:9, 9:16, 2:3, 3:2, 21:9, auto
sizeselectNo2K-1K, 1.5K, 2K
output_formatselectNopng-png, jpeg
backgroundselectNoauto-auto, transparent, opaque

Seedream 5.5 Layer Decomposition

seedream-5-5-layer-decompositionAvailable via APIFiles API needed

Image · Separate one image into a base image and transparent PNG layers with Seedream 5.5.

Credit pricing

The credit cost depends on your inputs and settings. Check the estimate before you start. The final cost is rounded up to whole credits.

Modes

image-to-layers

API input

Files API needed

Inputs

file

Prompt limit

2000 characters

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Reference files: Upload one JPEG, PNG or WebP up to 10 MB. An optional instruction accepts up to 2000 characters. Set a maximum of 2–100 output images, including the base; the default is 30. Auto size is budgeted at 2K.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
sizeselectNoauto-auto, 1K, 1.5K, 2K
output_formatselectNopng-png, jpeg
max_imagesnumberNo30min 2 / max 100 / step 1-
text_recognitionselectNofalse-false, true

Seedream 5.0 Lite

seedream-5-liteAvailable via APIText or URL; Files API for reference media

Image · ByteDance's lighter Seedream image model with shared pricing across text-to-image and image editing. It supports 8 aspect ratios, up to 14 reference images, and currently costs 6 credits per run.

Credit pricing

Credit rate: 5.5 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

2995 characters

Reference media

image · 14 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "seedream-5-lite",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspectRatio": "1:1",
    "quality": "basic",
    "output_format": "png",
    "nsfw_checker": "true"
  }
}

Reference rule

Reference image: Upload up to 14 image files for image projects.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspectRatioselectNo1:1-1:1, 4:3, 3:4, 16:9, 9:16, 2:3, 3:2, 21:9
qualityselectNobasic-basic, high, ultra
output_formatselectNopng-png, jpeg
nsfw_checkerselectNotrue-false, true

Seedream 5.0 Pro

seedream-5-proAvailable via APIText or URL; Files API for reference media

Image · Seedream 5.0 Pro is Rivya's multi-reference image model for campaign frames, editorial scenes, and product compositions, with text-to-image, image-to-image, 1K or 2K output, and up to 10 reference images.

Credit pricing

The credit cost depends on your inputs and settings. Check the estimate before you start. The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image · 10 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "seedream-5-pro",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspect_ratio": "1:1",
    "quality": "basic",
    "output_format": "png",
    "nsfw_checker": "false"
  }
}

Reference rule

Reference images: Upload up to 10 JPEG, PNG, or WebP images, no more than 10 MB each.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNo1:1-1:1, 4:3, 3:4, 16:9, 9:16, 2:3, 3:2, 21:9
qualityselectNobasic-basic, high
output_formatselectNopng-png, jpeg
nsfw_checkerselectNofalse-false, true

Seedream 5.0 Pro Layer Decomposition

seedream-5-pro-layer-decompositionAvailable via APIFiles API needed

Image · Seedream 5.0 Pro Layer Decomposition separates one source image into a base image and reusable PNG layers, with optional prompt or bounding-box guidance and Auto, 1K, 1.5K, or 2K sizing.

Credit pricing

Credit rate: 7–14 per image The final cost is rounded up to whole credits.

Modes

image-to-layers

API input

Files API needed

Inputs

file

Prompt limit

5000 characters

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Source image: Upload exactly one PNG/JPEG up to 30 MB; HEIC and HEIF are not supported.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
sizeselectNoauto-auto, 1K, 1.5K, 2K
output_formatselectNojpeg-jpeg, png

Seedream 5.0 Flash

seedream-5-flashAvailable via APIText or URL; Files API for reference media

Image · Create or edit an image at 1K, 1.5K or 2K in PNG or JPEG format. Add up to 10 JPEG, PNG, or WebP reference images, up to 30 MB each. Prompts must contain 4–5000 characters.

Credit pricing

Credit rate: 3.24 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image · 10 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "seedream-5-flash",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspect_ratio": "1:1",
    "size": "1K",
    "output_format": "png"
  }
}

Reference rule

Reference images: Create or edit an image at 1K, 1.5K or 2K in PNG or JPEG format. Add up to 10 JPEG, PNG, or WebP reference images, up to 30 MB each. Prompts must contain 4–5000 characters.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNo1:1-1:1, 4:3, 3:4, 16:9, 9:16, 2:3, 3:2, 21:9
sizeselectNo1K-1K, 1.5K, 2K
output_formatselectNopng-png, jpeg

Seedream 5.0 Flash Layer Decomposition

seedream-5-flash-layer-decompositionAvailable via APIFiles API needed

Image · Separate elements from one PNG or JPEG image up to 30 MB. Choose 1K, 1.5K or 2K. A prompt is optional; when provided it must contain 4–5000 characters.

Credit pricing

Credit rate: 3.24 per image The final cost is rounded up to whole credits.

Modes

image-to-layers

API input

Files API needed

Inputs

file

Prompt limit

5000 characters

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Reference images: Separate elements from one PNG or JPEG image up to 30 MB. Choose 1K, 1.5K or 2K. A prompt is optional; when provided it must contain 4–5000 characters.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
sizeselectNo1K-1K, 1.5K, 2K
output_formatselectNopng-png, jpeg

Seedream 4.5

seedream-4-5Available via APIText or URL; Files API for reference media

Image · ByteDance's high-end image model with 2K/4K quality tiers, 8 aspect ratios, and up to 14 reference images. Known for cinematic color grading and rich texture detail in fashion and lifestyle visuals.

Credit pricing

Credit rate: 6.5 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

3000 characters

Reference media

image · 14 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "seedream-4-5",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspectRatio": "1:1",
    "quality": "basic"
  }
}

Reference rule

Reference image: Upload up to 14 image files for image projects.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspectRatioselectNo1:1-1:1, 4:3, 3:4, 16:9, 9:16, 2:3, 3:2, 21:9
qualityselectNobasic-basic, high

Seedream 4.0

seedream-4-0Available via APIText or URL; Files API for reference media

Image · Create or edit up to 6 images in one request, at 1K, 2K or 4K. Add up to 10 JPEG, PNG or WebP references, up to 10 MB each. The reserve covers the maximum image count you request.

Credit pricing

Credit rate: 5 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image · 10 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "seedream-4-0",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "image_size": "square_hd",
    "image_resolution": "1K",
    "max_images": 1,
    "seed": 0,
    "nsfw_checker": "false"
  }
}

Reference rule

Reference images: Create or edit up to 6 images in one request, at 1K, 2K or 4K. Add up to 10 JPEG, PNG or WebP references, up to 10 MB each. The reserve covers the maximum image count you request.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
image_sizeselectNosquare_hd-square, square_hd, portrait_4_3, portrait_3_2, portrait_16_9, landscape_4_3, landscape_3_2, landscape_16_9, landscape_21_9
image_resolutionselectNo1K-1K, 2K, 4K
max_imagesnumberNo1min 1 / max 6 / step 1-
seednumberNo-min 0 / max 1000000 / step 1-
nsfw_checkerselectNofalse-false, true

Seedream 3.0

seedream-3-0Available via APIText input ready

Image · Create one image from a prompt of up to 5000 characters. Choose a size preset, guidance strength and optional seed.

Credit pricing

Credit rate: 3.5 per image The final cost is rounded up to whole credits.

Modes

text-to-image

API input

Text input ready

Inputs

text

Prompt limit

5000 characters

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "seedream-3-0",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "image_size": "square_hd",
    "guidance_scale": 2.5,
    "seed": 0
  }
}

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Parameters

keyTypeRequiredDefaultRangeOptions
image_sizeselectNosquare_hd-square, square_hd, portrait_4_3, portrait_16_9, landscape_4_3, landscape_16_9
guidance_scalenumberNo2.5min 1 / max 10 / step 0.1-
seednumberNo-min 0 / max 2147483647-

Grok Imagine Image 2.0

grok-imagine-image-2-0Available via APIText or URL; Files API for reference media

Image · Grok Imagine Image 2.0 supports single-image text generation and standard image editing with one to five signed reference images, five fixed aspect ratios, and an automatic ratio option for edit requests.

Credit pricing

Credit rate: 4 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

8000 characters

Reference media

image · 5 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "grok-imagine-image-2-0",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspect_ratio": "1:1"
  }
}

Reference rule

Reference images: Upload one to five JPEG, PNG, or WebP images, no more than 10 MB each. Signed upload metadata is required for every reference.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNo1:1-1:1, 2:3, 3:2, 16:9, 9:16, auto

Grok Imagine

grok-imagineAvailable via APIText or URL; Files API for reference media

Image · xAI's image model with strong creative interpretation and 5 aspect ratios. Standard text requests return 2 images, while quality requests return 4.

Credit pricing

Credit rate: 4–5 per request The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "grok-imagine",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspect_ratio": "1:1",
    "nsfw_checker": "false",
    "enable_pro": "false"
  }
}

Reference rule

Reference image: Upload up to 1 image file for image projects.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNo1:1-1:1, 2:3, 3:2, 16:9, 9:16
nsfw_checkerselectNofalse-false, true
enable_proselectNofalse-false, true

Flux 3 Klein

flux-3Available via APIText or URL; Files API for reference media

Image · Flux 3 Klein — Create one image from a prompt or reference images.

Credit pricing

Credit rate: 1.6–6.5 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image · 10 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "flux-3",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspect_ratio": "1:1",
    "resolution": "1K",
    "output_format": "png"
  }
}

Reference rule

Reference image: Up to 10 JPEG, PNG, or WebP reference images, each no larger than 30 MB.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNo1:1-1:1, 4:3, 3:4, 3:2, 2:3, 16:9, 9:16, 21:9, 9:21
resolutionselectNo1K-1K, 2K, 4K
output_formatselectNopng-png, jpeg, webp

Flux 2 Pro

flux-2-proAvailable via APIText or URL; Files API for reference media

Image · Create images from text, or guide an edit with image references.

Credit pricing

Credit rate: 5–7 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image · 8 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "flux-2-pro",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspectRatio": "1:1",
    "resolution": "1K",
    "nsfw_checker": "false"
  }
}

Reference rule

Reference images: Add up to 8 JPEG, PNG, or WebP images, each no larger than 10 MB. Explain in the prompt what you want to keep or change.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspectRatioselectNo1:1-1:1, 4:3, 3:4, 16:9, 9:16, 3:2, 2:3
resolutionselectNo1K-1K, 2K
nsfw_checkerselectNofalse-false, true

Flux 2 Flex

flux-2-flexAvailable via APIText or URL; Files API for reference media

Image · Generate images from text, or use references to try a new composition or style.

Credit pricing

Credit rate: 14–24 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image · 8 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "flux-2-flex",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspectRatio": "1:1",
    "resolution": "1K",
    "nsfw_checker": "false"
  }
}

Reference rule

Reference images: Use JPEG, PNG, or WebP files, up to 8 images and 10 MB per file. Tell the model what each reference is for.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspectRatioselectNo1:1-1:1, 4:3, 3:4, 16:9, 9:16, 3:2, 2:3
resolutionselectNo1K-1K, 2K
nsfw_checkerselectNofalse-false, true

Flux Kontext Max

flux-kontext-maxAvailable via APIText or URL; Files API for reference media

Image · Flux Kontext Max is a previous-generation Black Forest Labs model for generating images from text and editing a source image. BFL now recommends FLUX.2 for new projects.

Credit pricing

Credit rate: 8 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

-

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "flux-kontext-max",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspectRatio": "16:9",
    "outputFormat": "jpeg",
    "enableTranslation": "true",
    "promptUpsampling": "false",
    "safetyTolerance": 2,
    "watermark": "Optional watermark identifier"
  }
}

Reference rule

Reference image: Add up to 1 source image for editing. JPEG, PNG, or WebP; up to 10 MiB.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspectRatioselectNo16:9-21:9, 16:9, 4:3, 1:1, 3:4, 9:16
outputFormatselectNojpeg-jpeg, png
enableTranslationselectNotrue-true, false
promptUpsamplingselectNofalse-false, true
safetyTolerancenumberNo2min 0 / max 6-
watermarktextNo---

Flux Kontext Pro

flux-kontext-proAvailable via APIText or URL; Files API for reference media

Image · Flux Kontext Pro creates images from text or edits a reference image on Rivya. Black Forest Labs lists FLUX.1 Kontext [pro] as a previous-generation model and recommends FLUX.2 for new projects.

Credit pricing

Credit rate: 4 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

-

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "flux-kontext-pro",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspectRatio": "16:9",
    "outputFormat": "jpeg",
    "enableTranslation": "true",
    "promptUpsampling": "false",
    "safetyTolerance": 2,
    "watermark": "Optional watermark identifier"
  }
}

Reference rule

Reference image: Upload a JPEG, PNG, or WebP image to edit, up to 10 MB. Sign in before uploading.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspectRatioselectNo16:9-21:9, 16:9, 4:3, 1:1, 3:4, 9:16
outputFormatselectNojpeg-jpeg, png
enableTranslationselectNotrue-true, false
promptUpsamplingselectNofalse-false, true
safetyTolerancenumberNo2min 0 / max 6-
watermarktextNo---

Qwen Image 3.0

qwen-image-3Available via APIText or URL; Files API for reference media

Image · Qwen Image 3.0 is Rivya's configurable Qwen image model for Chinese or English prompts, reference-led layouts, and reproducible visual tests with 1K or 2K output, prompt expansion, negative prompts, and seed control.

Credit pricing

The credit cost depends on your inputs and settings. Check the estimate before you start. The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

800 characters

Reference media

image · 3 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "qwen-image-3",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "resolution": "1K",
    "image_size": "16:9",
    "output_format": "png",
    "prompt_extend": "true",
    "nsfw_checker": "false",
    "negative_prompt": "Optional elements or failure modes to avoid, up to 5,000 characters.",
    "seed": 1
  }
}

Reference rule

Reference images: Upload up to three JPEG, PNG, WebP, BMP, GIF, or TIFF images, no more than 10 MB each.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
resolutionselectYes1K-1K, 2K
image_sizeselectNo16:9-1:1, 3:2, 2:3, 4:3, 3:4, 16:9, 9:16, 21:9
output_formatselectNopng-png, jpeg
prompt_extendselectNotrue-true, false
nsfw_checkerselectNofalse-false, true
negative_prompttextNo---
seednumberNo1min 0 / max 2147483647 / step 1-

Qwen Image 3.0 Pro

qwen-image-3-proAvailable via APIText or URL; Files API for reference media

Image · Qwen Image 3.0 Pro is Rivya's separate Pro-tier Qwen image model for polished campaign key visuals, product scenes, and poster concepts, with Chinese or English prompts, up to three references, and distinct 1K and 2K credit tiers.

Credit pricing

The credit cost depends on your inputs and settings. Check the estimate before you start. The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

800 characters

Reference media

image · 3 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "qwen-image-3-pro",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "resolution": "1K",
    "image_size": "16:9",
    "output_format": "png",
    "prompt_extend": "true",
    "nsfw_checker": "false",
    "negative_prompt": "Optional elements or failure modes to avoid, up to 5,000 characters.",
    "seed": 1
  }
}

Reference rule

Reference images: Upload up to three JPEG, PNG, WebP, BMP, GIF, or TIFF images, no more than 10 MB each.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
resolutionselectYes1K-1K, 2K
image_sizeselectNo16:9-1:1, 3:2, 2:3, 4:3, 3:4, 16:9, 9:16, 21:9
output_formatselectNopng-png, jpeg
prompt_extendselectNotrue-true, false
nsfw_checkerselectNofalse-false, true
negative_prompttextNo---
seednumberNo1min 0 / max 2147483647 / step 1-

Qwen Image 2.1

qwen2-1-imageAvailable via APIText or URL; Files API for reference media

Image · Create or edit one image at 1K or 2K. Add up to 10 JPEG, PNG, or WebP references, up to 30 MB and 25 megapixels each. Transparent backgrounds require PNG or WebP. Masks are unavailable here.

Credit pricing

Credit rate: 4–8 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image · 10 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "qwen2-1-image",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspect_ratio": "1:1",
    "resolution": "1K",
    "output_format": "png",
    "background": "opaque",
    "enhance_prompt": "true",
    "seed": 0
  }
}

Reference rule

Reference images: Create or edit one image at 1K or 2K. Add up to 10 JPEG, PNG, or WebP references, up to 30 MB and 25 megapixels each. Transparent backgrounds require PNG or WebP. Masks are unavailable here.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNo1:1-1:1, 4:3, 3:4, 3:2, 2:3, 16:9, 9:16, 21:9, 9:21
resolutionselectNo1K-1K, 2K
output_formatselectNopng-png, webp, jpeg
backgroundselectNoopaque-opaque, transparent
enhance_promptselectNotrue-true, false
seednumberNo-min 0 / max 2147483647 / step 1-

Qwen2 Image

qwen2-imageAvailable via APIText or URL; Files API for reference media

Image · Alibaba's Qwen2 image model is currently integrated on Rivya as one fixed-price image project. It safely covers text-to-image and image-to-image with the shared aspect-ratio subset both public docs expose, plus PNG/JPEG output, seed reuse, and a simple NSFW switch.

Credit pricing

Credit rate: 5.6 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

800 characters

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "qwen2-image",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "image_size": "16:9",
    "output_format": "png",
    "seed": 0,
    "nsfw_checker": "false"
  }
}

Reference rule

Reference image: Upload up to 1 image file for image projects.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
image_sizeselectNo16:9-1:1, 2:3, 3:2, 3:4, 4:3, 9:16, 16:9, 21:9
output_formatselectNopng-png, jpeg
seednumberNo-min 0-
nsfw_checkerselectNofalse-false, true

Qwen Image

qwen-imageAvailable via APIText or URL; Files API for reference media

Image · Alibaba Qwen family's image model with HD presets (Square, Portrait, Landscape) and PNG/JPEG output. Strong at Chinese-language prompts and culturally nuanced visual generation.

Credit pricing

Credit rate: 4 per megapixel of output The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "qwen-image",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "image_size": "square_hd",
    "output_format": "png"
  }
}

Reference rule

Reference image: Upload up to 1 image file for image projects.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
image_sizeselectNosquare_hd-square, square_hd, portrait_4_3, portrait_16_9, landscape_4_3, landscape_16_9
output_formatselectNopng-png, jpeg

Qwen Image Edit

qwen-image-editAvailable via APIFiles API needed

Image · On Rivya, edit one JPEG, PNG, or WebP image up to 10 MB and 16 megapixels using a prompt of up to 2000 characters. The result is one 1024 × 1024 PNG image.

Credit pricing

Credit rate: 5 per megapixel of output The final cost is rounded up to whole credits.

Modes

image-to-image

API input

Files API needed

Inputs

file

Prompt limit

2000 characters

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Reference images: On Rivya, edit one JPEG, PNG, or WebP image up to 10 MB and 16 megapixels using a prompt of up to 2000 characters. The result is one 1024 × 1024 PNG image.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
image_sizeselectNosquare_hd-square_hd
output_formatselectNopng-png

Midjourney

midjourneyThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Image · Midjourney is an image-model family whose versions have different reference behavior. This historical Rivya entry is paused.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

image · 4 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

image: image

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Recraft Remove Background

recraft-remove-backgroundAvailable via APIFiles API needed

Image · Recraft's background-removal model on Rivya for isolating the subject from one existing image. Use it when the next step needs a transparent asset, a clean cutout, or a source image without the original background.

Credit pricing

Credit rate: 1 per image The final cost is rounded up to whole credits.

Modes

image-to-image

API input

Files API needed

Inputs

file

Prompt limit

-

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Source image: Upload exactly 1 JPG, PNG, or WebP image, up to 5 MB and 16 MP; each dimension must be 256–4096 px.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Recraft Crisp Upscale

recraft-crisp-upscaleAvailable via APIFiles API needed

Image · Recraft's light image-upscaling model on Rivya for low-cost sharpness and clarity boosts on one approved still. Use it when the chosen image only needs a cheap polish pass before export, not a heavier delivery-grade upscale.

Credit pricing

Credit rate: 0.5 per image The final cost is rounded up to whole credits.

Modes

image-to-image

API input

Files API needed

Inputs

file

Prompt limit

-

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Source image: Upload exactly 1 JPG, PNG, or WebP image, up to 10 MB.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Wan 2.7 Image Pro

wan-2-7-image-proAvailable via APIText or URL; Files API for reference media

Image · On Rivya, Wan 2.7 Image Pro currently supports text to image and image to image. It can also take up to 9 reference images. Fixed 12 credits per image; one image per request.

Credit pricing

Credit rate: 12 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image · 9 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "wan-2-7-image-pro",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspect_ratio": "1:1",
    "resolution": "2K",
    "watermark": "false",
    "seed": 0
  }
}

Reference rule

Reference image: Upload up to 9 image files for image editing projects. JPEG / PNG / WebP · ≤ 10 MB · ≥ 240 × 240 px.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNo1:1-1:1, 16:9, 4:3, 21:9, 3:4, 9:16, 8:1, 1:8
resolutionselectNo2K-1K, 2K
watermarkselectNofalse-false, true
seednumberNo-min 0 / max 2147483647 / step 1-

Wan 2.7 Image

wan-2-7-imageAvailable via APIText or URL; Files API for reference media

Image · On Rivya, Wan 2.7 Image currently supports text to image and image to image. It can also take up to 9 reference images. Fixed 5 credits per image; one image per request.

Credit pricing

Credit rate: 4.8 per image The final cost is rounded up to whole credits.

Modes

text-to-image, image-to-image

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image · 9 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "wan-2-7-image",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspect_ratio": "1:1",
    "resolution": "2K",
    "watermark": "false",
    "seed": 0
  }
}

Reference rule

Reference image: Upload up to 9 image files for image editing projects. JPEG / PNG / WebP · ≤ 10 MB · ≥ 240 × 240 px.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNo1:1-1:1, 16:9, 4:3, 21:9, 3:4, 9:16, 8:1, 1:8
resolutionselectNo2K-1K, 2K
watermarkselectNofalse-false, true
seednumberNo-min 0 / max 2147483647-

Google Imagen4 Ultra

google-imagen4-ultraAvailable via APIText input ready

Image · Google Imagen 4 Ultra creates images from text in Rivya. Set the prompt, negative prompt, aspect ratio and optional seed text; this workflow has no reference-image input or image-count control.

Credit pricing

Credit rate: 12 per image The final cost is rounded up to whole credits.

Modes

text-to-image

API input

Text input ready

Inputs

text

Prompt limit

5000 characters

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "google-imagen4-ultra",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspect_ratio": "1:1",
    "negative_prompt": "Optional elements to avoid",
    "seed": "Optional seed text"
  }
}

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNo1:1-1:1, 16:9, 9:16, 3:4, 4:3, auto
negative_prompttextNo-max 5000-
seedtextNo-max 500-

Google Imagen4

google-imagen4Available via APIText input ready

Image · Google Imagen4 is the standard Imagen text-to-image option on Rivya. Describe a scene in text, then choose an aspect ratio and optional negative prompt or seed.

Credit pricing

Credit rate: 8 per request The final cost is rounded up to whole credits.

Modes

text-to-image

API input

Text input ready

Inputs

text

Prompt limit

5000 characters

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "google-imagen4",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspect_ratio": "1:1",
    "negative_prompt": "Optional things to avoid",
    "seed": "Optional seed text"
  }
}

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNo1:1-1:1, 16:9, 9:16, 3:4, 4:3, auto
negative_prompttextNo-max 5000-
seedtextNo-max 500-

Google Imagen4 Fast

google-imagen4-fastAvailable via APIText input ready

Image · Google Imagen4 Fast is a Google text-to-image model offered through Rivya for landing-page visuals, blog art, social graphics, and product or poster concepts. Describe the subject and layout, then use the aspect ratio, negative prompt, and optional seed to prepare the request.

Credit pricing

Credit rate: 4 per request The final cost is rounded up to whole credits.

Modes

text-to-image

API input

Text input ready

Inputs

text

Prompt limit

5000 characters

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "google-imagen4-fast",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "aspect_ratio": "16:9",
    "negative_prompt": "Optional things to avoid",
    "seed": 1
  }
}

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Parameters

keyTypeRequiredDefaultRangeOptions
aspect_ratioselectNo16:9-1:1, 16:9, 9:16, 3:4, 4:3, auto
negative_prompttextNo-max 5000-
seednumberNo---

Topaz Image Upscaler

topaz-image-upscaleAvailable via APIFiles API needed

Image · Topaz's delivery-grade image upscaler on Rivya for approved stills that need a real size jump. Use it when the composition is already final and the remaining problem is export resolution, review size, or print readiness.

Credit pricing

Credit rate: 10–20 per image The final cost is rounded up to whole credits.

Modes

image-to-image

API input

Files API needed

Inputs

file

Prompt limit

-

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Source image: Exactly 1 JPEG, PNG, or WebP image up to 10 MB; output longest edge ≤ 4096 px. This is Rivya’s currently supported output range.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
upscale_factorselectNo2-1, 2, 4

Ideogram V3

ideogram-v3Available via APIText input ready

Image · Create images from text with Ideogram V3 on Rivya. Plan poster concepts, cover art, or social graphics, then check the lettering and composition before use.

Credit pricing

Credit rate: 3.5–10 per image The final cost is rounded up to whole credits.

Modes

text-to-image

API input

Text input ready

Inputs

text

Prompt limit

5000 characters

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "ideogram-v3",
  "prompt": "A product bottle on a light gray background, lit softly from the left.",
  "params": {
    "rendering_speed": "BALANCED",
    "style": "AUTO",
    "expand_prompt": "true",
    "image_size": "square_hd",
    "seed": 1,
    "negative_prompt": "Describe details to avoid"
  }
}

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Parameters

keyTypeRequiredDefaultRangeOptions
rendering_speedselectNoBALANCED-TURBO, BALANCED, QUALITY
styleselectNoAUTO-AUTO, GENERAL, REALISTIC, DESIGN
expand_promptselectNotrue-true, false
image_sizeselectNosquare_hd-square, square_hd, portrait_4_3, portrait_16_9, landscape_4_3, landscape_16_9
seednumberNo---
negative_prompttextNo-max 5000-

Ideogram V3 Reframe

ideogram-v3-reframeThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Image · Ideogram V3 Reframe expands an image canvas while keeping the source pixels in the center. This operation is paused on Rivya.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

image: image

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Ideogram V3 Remix

ideogram-v3-remixAvailable via APIFiles API needed

Image · Ideogram V3 Remix: TURBO 4 / BALANCED 7 / QUALITY 10 credits per image. Choose 1–4 outputs. The total is the per-image rate multiplied by the selected count, rounded up to whole credits.

Credit pricing

Credit rate: 3.5–10 per image The final cost is rounded up to whole credits.

Modes

image-to-image

API input

Files API needed

Inputs

file

Prompt limit

5000 characters

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Reference image: Upload exactly 1 image file for this project.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
rendering_speedselectNoBALANCED-TURBO, BALANCED, QUALITY
styleselectNoAUTO-AUTO, GENERAL, REALISTIC, DESIGN
expand_promptselectNotrue-true, false
image_sizeselectNosquare_hd-square, square_hd, portrait_4_3, portrait_16_9, landscape_4_3, landscape_16_9
num_imagesselectNo1-1, 2, 3, 4
seednumberNo---
strengthnumberNo0.8min 0.01 / max 1-
negative_prompttextNo-max 5000-

Ideogram Character

ideogram-characterAvailable via APIFiles API needed

Image · Character-consistency option for turning one approved character image into new scenes, outfits, and formats. TURBO 12 / BALANCED 18 / QUALITY 24 credits per image. Choose 1–4 outputs. The total is the per-image rate multiplied by the selected count, rounded up to whole credits.

Credit pricing

Credit rate: 12–24 per image The final cost is rounded up to whole credits.

Modes

image-to-image

API input

Files API needed

Inputs

file

Prompt limit

5000 characters

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Character reference image: Upload exactly 1 image file for this project.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
rendering_speedselectNoBALANCED-TURBO, BALANCED, QUALITY
styleselectNoAUTO-AUTO, REALISTIC, FICTION
expand_promptselectNotrue-true, false
image_sizeselectNosquare_hd-square, square_hd, portrait_4_3, portrait_16_9, landscape_4_3, landscape_16_9
num_imagesselectNo1-1, 2, 3, 4
seednumberNo---
negative_prompttextNo-max 5000-

Rivya Image Upscale

image-upscaleAvailable via APIFiles API needed

Image · Enhance or enlarge one image by 1×, 2×, 4× or 8× while keeping its aspect ratio.

Credit pricing

Credit rate: 1.5–12 per image The final cost is rounded up to whole credits.

Modes

image-to-image

API input

Files API needed

Inputs

file

Prompt limit

-

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Reference files: Upload one JPEG, PNG, WebP, HEIC, BMP or GIF up to 20 MB. GIF uses the first frame. Output is limited to 67,108,864 total pixels and 16,384 pixels on either side.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
upscale_factorselectNo2-1, 2, 4, 8

MiniMax H3 Image

minimax-h3-imageThis model is coming soon and cannot currently be called through Public API v1.This model is coming soon and cannot currently be called through Public API v1.

Image · MiniMax H3 Image preview: separate text-to-image and reference-image editing operations, with generation still unavailable on Rivya.

Credit pricing

This model is coming soon and cannot currently be called through Public API v1.

Modes

API input

This model is coming soon and cannot currently be called through Public API v1.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model is coming soon and cannot currently be called through Public API v1.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Seedance 2.5

seedance-2-5Available via APIText or URL; Files API for reference media

Video · Create Seedance 2.5 videos from text, a start frame with an optional end frame, or image, video, and audio references. Choose 480p, 720p, or 1080p, a duration from 4 to 30 seconds or automatic duration, and an MP4 or MOV format request.

Credit pricing

Credit rate: 17–158 per second of total video duration (input + output) The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video, reference-to-video, video-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

20480 characters

Reference media

image, video, audio · 50 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "seedance-2-5",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "seedance_scene": "text",
    "resolution": "720p",
    "aspect_ratio": "adaptive",
    "duration": 5,
    "output_format": "mp4",
    "generate_audio": "true",
    "return_last_frame": "false",
    "web_search": "false",
    "nsfw_checker": "true"
  }
}

Reference rule

Guide frames or image, video, and audio references: Frames accepts 1 or 2 JPEG/PNG/WebP images up to 30 MB each. Reference accepts up to 30 images, 10 MP4/MOV videos up to 95 MB and 2–30 seconds each, and 10 MP3/WAV audio files up to 15 MB and 2–30 seconds each. Video and audio totals are each capped at 30 seconds, with no more than 50 files overall.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
omni_reference_task_typeselectNoauto-auto, reference, edit, extend
seedance_sceneselectNotext-text, frames, reference
resolutionselectNo720p-480p, 720p, 1080p
aspect_ratioselectNoadaptive-1:1, 4:3, 3:4, 16:9, 9:16, 21:9, adaptive
durationnumberNo5min -1 / max 30 / step 1-
output_formatselectNomp4-mp4, mov
generate_audioselectNotrue-true, false
return_last_frameselectNofalse-false, true
web_searchselectNofalse-false, true
nsfw_checkerselectNotrue-false, true

Seedance 2.0

seedance-2Available via APIText or URL; Files API for reference media

Video · Generate video from text, a first frame, or media references.

Credit pricing

Credit rate: 41–208 per second of total video duration (input + output) The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video, reference-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

20000 characters

Reference media

image, video, audio · 15 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "seedance-2",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "seedance_scene": "text",
    "resolution": "720p",
    "aspect_ratio": "16:9",
    "duration": 5,
    "generate_audio": "true",
    "return_last_frame": "false",
    "web_search": "false"
  }
}

Reference rule

Frames and media references: Frame mode needs a first image and may include a last image. Reference mode allows up to 9 images, 3 videos, and 3 audio clips. Use JPEG, PNG, WebP, BMP, TIFF, or GIF images up to 30 MB each; MP4 or MOV video up to 50 MB each; and MP3 or WAV audio up to 15 MB each. Each video or audio clip must be between 2 and 15 seconds. Video references together are limited to 15 seconds, and audio has a separate 15-second total. Video references require 1080p or 4K output. Rivya also checks image dimensions and video frame size and frame rate. Upload through Rivya so the request includes the required signed metadata.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
seedance_sceneselectNotext-text, frames, reference
resolutionselectNo720p-720p, 1080p, 4K
aspect_ratioselectNo16:9-1:1, 4:3, 3:4, 16:9, 9:16, 21:9, adaptive
durationnumberNo5min 4 / max 15-
generate_audioselectNotrue-true, false
return_last_frameselectNofalse-false, true
web_searchselectNofalse-false, true

Seedance 2.0 Fast

seedance-2-fastAvailable via APIText or URL; Files API for reference media

Video · ByteDance's Seedance 2.0 Fast video model for text prompts, a first frame with an optional last frame, or image, video and audio references. Choose 480p or 720p, set a 4–15-second duration, and request audio or a returned last frame.

Credit pricing

Credit rate: 6.8–24.8 per second of total video duration (input + output) The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video, reference-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

20000 characters

Reference media

image, video, audio · 15 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "seedance-2-fast",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "seedance_scene": "text",
    "resolution": "720p",
    "aspect_ratio": "16:9",
    "duration": 5,
    "generate_audio": "true",
    "return_last_frame": "false",
    "web_search": "false",
    "nsfw_checker": "false"
  }
}

Reference rule

Multimodal reference: Frames: upload a first image and, optionally, a last image. References: up to 9 JPEG/PNG/WebP/BMP/TIFF/GIF images at 30 MiB each or less, 3 MP4/MOV videos at 50 MiB each or less, and 3 MP3/WAV audio clips at 15 MiB each or less. Each video or audio clip must be 2–15 seconds; video and audio each have a separate 15-second total limit. Uploads must also pass file and metadata checks.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
seedance_sceneselectNotext-text, frames, reference
resolutionselectNo720p-480p, 720p
aspect_ratioselectNo16:9-1:1, 4:3, 3:4, 16:9, 9:16, 21:9, adaptive
durationnumberNo5min 4 / max 15-
generate_audioselectNotrue-true, false
return_last_frameselectNofalse-false, true
web_searchselectNofalse-false, true
nsfw_checkerselectNofalse-false, true

Seedance 2.0 Mini

seedance-2-miniAvailable via APIText or URL; Files API for reference media

Video · Seedance 2.0 Mini creates 480p or 720p videos lasting 4–15 seconds. Start with text, guide the clip with a first frame or a first-and-last-frame pair, or combine image, video and audio references. Choose one input mode for each generation, with optional generated audio.

Credit pricing

Credit rate: 2.4–8.2 per second of total video duration (input + output) The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video, reference-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

20000 characters

Reference media

image, video, audio · 15 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "seedance-2-mini",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "seedance_scene": "text",
    "resolution": "720p",
    "aspect_ratio": "16:9",
    "duration": 5,
    "generate_audio": "true",
    "web_search": "false",
    "nsfw_checker": "true"
  }
}

Reference rule

Guide frames or multimodal references: For Frames, upload one or two JPEG/PNG/WebP/GIF images up to 30 MB each. Reference accepts up to 9 images in those formats, 3 MP4/MOV videos up to 50 MB each, and 3 MP3/WAV audio files up to 15 MB each. Each video or audio file must last 2–15 seconds; video and audio each have a separate 15-second total limit.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
seedance_sceneselectNotext-text, frames, reference
resolutionselectNo720p-480p, 720p
aspect_ratioselectNo16:9-1:1, 4:3, 3:4, 16:9, 9:16, 21:9, adaptive
durationnumberNo5min 4 / max 15 / step 1-
generate_audioselectNotrue-true, false
web_searchselectNofalse-false, true
nsfw_checkerselectNotrue-false, true

Seedance 1.5 Pro

seedance-1-5-proAvailable via APIText or URL; Files API for reference media

Video · Create video from a prompt or up to 2 reference images with Seedance 1.5 Pro. Choose 480p, 720p, or 1080p, set a duration of 4–12 seconds, and turn generated audio on or off.

Credit pricing

Credit rate: 1.75–15 per second of output The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

2500 characters

Reference media

image · 2 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "seedance-1-5-pro",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "resolution": "720p",
    "duration": 8,
    "generate_audio": "false",
    "aspectRatio": "1:1",
    "fixed_lens": "false",
    "nsfw_checker": "false"
  }
}

Reference rule

Reference images: Upload up to 2 JPEG, PNG, or WebP images, each at most 10 MB, and describe the motion in your prompt.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
resolutionselectNo720p-480p, 720p, 1080p
durationnumberNo8min 4 / max 12 / step 1-
generate_audioselectNofalse-false, true
aspectRatioselectNo1:1-1:1, 4:3, 3:4, 16:9, 9:16, 21:9
fixed_lensselectNofalse-false, true
nsfw_checkerselectNofalse-false, true

Seedance 1.0 Pro

seedance-1-0-proAvailable via APIText or URL; Files API for reference media

Video · Create a 5- or 10-second video at 480p, 720p or 1080p from text or one image. Each JPEG, PNG or WebP image must be at most 10 MB.

Credit pricing

Credit rate: 14–140 per video The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

10000 characters

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "seedance-1-0-pro",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "resolution": "720p",
    "duration": "5",
    "aspect_ratio": "16:9",
    "camera_fixed": "false",
    "seed": -1,
    "enable_safety_checker": "true"
  }
}

Reference rule

Reference images: Create a 5- or 10-second video at 480p, 720p or 1080p from text or one image. Each JPEG, PNG or WebP image must be at most 10 MB.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
resolutionselectNo720p-480p, 720p, 1080p
durationselectNo5-5, 10
aspect_ratioselectNo16:9-21:9, 16:9, 4:3, 1:1, 3:4, 9:16
camera_fixedselectNofalse-false, true
seednumberNo-1min -1 / max 2147483647-
enable_safety_checkerselectNotrue-true, false

Seedance 1.0 Pro Fast

seedance-1-0-pro-fastAvailable via APIFiles API needed

Video · Create a 5- or 10-second video at 720p or 1080p from one required JPEG, PNG or WebP image, up to 10 MB.

Credit pricing

Credit rate: 16–72 per video The final cost is rounded up to whole credits.

Modes

image-to-video

API input

Files API needed

Inputs

file

Prompt limit

10000 characters

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Reference images: Create a 5- or 10-second video at 720p or 1080p from one required JPEG, PNG or WebP image, up to 10 MB.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
resolutionselectNo720p-720p, 1080p
durationselectNo5-5, 10

Seedance 1.0 Lite

seedance-1-0-liteAvailable via APIText or URL; Files API for reference media

Video · Create a 5- or 10-second video at 480p, 720p or 1080p from text or one image. Lite also accepts an optional end-frame image. Each JPEG, PNG or WebP image must be at most 10 MB.

Credit pricing

Credit rate: 10–100 per video The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

10000 characters

Reference media

image · 2 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "seedance-1-0-lite",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "resolution": "720p",
    "duration": "5",
    "aspect_ratio": "16:9",
    "camera_fixed": "false",
    "seed": -1,
    "enable_safety_checker": "true"
  }
}

Reference rule

Reference images: Create a 5- or 10-second video at 480p, 720p or 1080p from text or one image. Lite also accepts an optional end-frame image. Each JPEG, PNG or WebP image must be at most 10 MB.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
resolutionselectNo720p-480p, 720p, 1080p
durationselectNo5-5, 10
aspect_ratioselectNo16:9-16:9, 4:3, 1:1, 3:4, 9:16, 9:21
camera_fixedselectNofalse-false, true
seednumberNo-1min -1 / max 2147483647-
enable_safety_checkerselectNotrue-true, false

HappyHorse 1.1

happyhorse-1-1Available via APIText or URL; Files API for reference media

Video · HappyHorse 1.1 is a flexible AI video model for prompt-only clips, single-image animation, and multi-image reference video at 720p or 1080p, with 3 to 15 seconds of output.

Credit pricing

Credit rate: 22.5–29 per second of output The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video, reference-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

4999 characters

Reference media

image · 9 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "happyhorse-1-1",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "resolution": "1080p",
    "aspect_ratio": "16:9",
    "duration": 5
  }
}

Reference rule

First frame or reference images: Upload one JPEG, PNG, or WebP image for first-frame animation, or 2 to 9 images for reference-led video. Each file can be up to 20 MB.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
resolutionselectNo1080p-720p, 1080p
aspect_ratioselectNo16:9-16:9, 9:16, 1:1, 4:3, 3:4, 4:5, 5:4, 9:21, 21:9
durationnumberNo5min 3 / max 15 / step 1-

HappyHorse 1.0

happyhorse-1-0Available via APIText or URL; Files API for reference media

Video · HappyHorse 1.0 on Rivya supports text-to-video, single-image animation, and multi-image reference video from one model page.

Credit pricing

Credit rate: 28–48 per second of output The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video, reference-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image · 9 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "happyhorse-1-0",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "resolution": "1080p",
    "aspect_ratio": "16:9",
    "duration": 5,
    "seed": 0
  }
}

Reference rule

Reference media: Use exactly 1 image for image-to-video or 2 to 9 images for reference video. Images must be JPEG, PNG, or WebP and no larger than 10 MB; single-image mode requires both dimensions to be at least 300 px, while reference mode requires the shortest side to be at least 400 px.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
resolutionselectNo1080p-720p, 1080p
aspect_ratioselectNo16:9-16:9, 9:16, 1:1, 4:3, 3:4
durationnumberNo5min 3 / max 15-
seednumberNo0min 0 / max 2147483647-

Wan 3.0 Video

wan-3-0-videoAvailable via APIText or URL; Files API for reference media

Video · Wan 3.0 Video combines Standard and Prime tiers with text-to-video, first/last-frame guidance, and signed multimodal references. It supports 480P, 720P, or 1080P output, explicit 2-30-second clips or intelligent duration, and reference bundles containing images, video, and audio.

Credit pricing

Credit rate: 8–50.4 per second of total video duration (input + output) The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video, reference-to-video, video-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

20000 characters

Reference media

image, video, audio · 20 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "wan-3-0-video",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "seedance_scene": "text",
    "variant": "standard",
    "resolution": "1080P",
    "aspect_ratio": "adaptive",
    "duration": 5,
    "audio": "true",
    "seed": 0
  }
}

Reference rule

Guide frames or signed multimodal references: Frames accepts a required first image and one optional last image. Reference accepts up to 10 JPEG, non-transparent PNG, WebP, or BMP images at 20 MB each, 5 MP4/MOV videos at 95 MB and 1-15 seconds each, and 5 MP3/WAV audio clips at 15 MB and 1-15 seconds each; video and audio totals are each capped at 15 seconds.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
seedance_sceneselectNotext-text, frames, reference
variantselectNostandard-standard, prime
resolutionselectNo1080P-480P, 720P, 1080P
aspect_ratioselectNoadaptive-adaptive, 16:9, 4:3, 1:1, 3:4, 9:16
durationnumberNo5min -1 / max 30 / step 1-
audioselectNotrue-true, false
seednumberNo-min 0 / max 2147483647 / step 1-

Wan 2.7 Video

wan-2-7-videoAvailable via APIText or URL; Files API for reference media

Video · Generate short videos from text or images, or edit an existing clip.

Credit pricing

Credit rate: 16–24 per second of output The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video, video-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image, video · 2 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "wan-2-7-video",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "resolution": "720p",
    "duration": 5,
    "aspect_ratio": "16:9",
    "negative_prompt": "Elements to avoid, such as blur or unwanted objects",
    "prompt_extend": "true",
    "watermark": "false",
    "audio_setting": "auto",
    "seed": 0
  }
}

Reference rule

Reference media: Add a first frame and, optionally, a last frame. For an edit, add a source video with an optional image reference. Images must be JPEG, PNG, or WebP and no larger than 30 MB; videos must be MP4 or MOV and no larger than 30 MB.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
resolutionselectNo720p-720p, 1080p
durationnumberNo5min 0 / max 15-
aspect_ratioselectNo16:9-16:9, 9:16, 1:1, 4:3, 3:4
negative_prompttextNo-max 500-
prompt_extendselectNotrue-true, false
watermarkselectNofalse-false, true
audio_settingselectNoauto-auto, origin
seednumberNo-min 0 / max 2147483647-

Wan 2.6

wan-2-6Available via APIText or URL; Files API for reference media

Video · Create a short video from text, one image, or one to three source clips. A written prompt is required in every mode, and image and video references cannot be mixed.

Credit pricing

Credit rate: 70–315 per video The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video, video-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image, video · 3 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "wan-2-6",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "resolution": "1080p",
    "duration": "5",
    "multi_shots": "false",
    "nsfw_checker": "true"
  }
}

Reference rule

Reference files: Choose one JPEG, PNG, or WebP image with both dimensions at least 256 pixels, or one to three MP4 or MOV videos. Each file must be 10 MiB (10,485,760 bytes) or smaller. Images and videos cannot be combined.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
resolutionselectNo1080p-720p, 1080p
durationselectNo5-5, 10, 15
multi_shotsselectNofalse-false, true
nsfw_checkerselectNotrue-false, true

Wan 2.6 Flash

wan-2-6-flashAvailable via APIFiles API needed

Video · Create video from 1 image or 1–3 reference videos, up to 10 MB each. Choose 720p or 1080p and optional generated audio. Image input supports 5/10/15 seconds; video input supports 5/10 seconds.

Credit pricing

Credit rate: 20–180 per video The final cost is rounded up to whole credits.

Modes

image-to-video, video-to-video

API input

Files API needed

Inputs

file

Prompt limit

1500 characters

Reference media

image, video · 3 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Reference image or videos: Create video from 1 image or 1–3 reference videos, up to 10 MB each. Choose 720p or 1080p and optional generated audio. Image input supports 5/10/15 seconds; video input supports 5/10 seconds.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
resolutionselectNo1080p-720p, 1080p
durationselectNo5-5, 10, 15
audioselectNofalse-false, true
multi_shotsselectNofalse-false, true

Wan 2.5 Video

wan-2-5-videoAvailable via APIText or URL; Files API for reference media

Video · Generate 5- or 10-second videos from a prompt or one image, with 720p and 1080p options. A written prompt is required in both modes.

Credit pricing

Credit rate: 60–200 per video The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

800 characters

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "wan-2-5-video",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "resolution": "720p",
    "duration": "5",
    "aspect_ratio": "16:9",
    "enable_prompt_expansion": "false",
    "seed": 0,
    "negative_prompt": "Optional: describe what to avoid (up to 500 characters)"
  }
}

Reference rule

Reference image: Sign in to upload one JPEG, PNG, or WebP image, up to 10 MiB (10,485,760 bytes). A written prompt is also required.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
resolutionselectNo720p-720p, 1080p
durationselectNo5-5, 10
aspect_ratioselectNo16:9-16:9, 9:16, 1:1
enable_prompt_expansionselectNofalse-false, true
seednumberNo-min 0 / max 2147483647-
negative_prompttextNo-max 500-

Wan 2.2 A14B Turbo

wan-2-2-a14b-turboAvailable via APIText or URL; Files API for reference media

Video · Text and image input support 480p or 720p (40 or 80 credits for 5 seconds). Image plus audio also supports 580p; its 480p / 580p / 720p rates are 12 / 18 / 24 credits per output second.

Credit pricing

Credit rate: 12–80 per second of output The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video, speech-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image, audio · 2 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "wan-2-2-a14b-turbo",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "resolution": "720p",
    "aspect_ratio": "16:9",
    "enable_prompt_expansion": "false",
    "seed": 0,
    "acceleration": "none",
    "num_frames": 80,
    "frames_per_second": 16,
    "negative_prompt": "Optional exclusions or failure modes to avoid.",
    "num_inference_steps": 27,
    "guidance_scale": 3.5,
    "shift": 5,
    "enable_safety_checker": "true"
  }
}

Reference rule

Reference media: Upload up to 2 assets. Text mode needs none, image-to-video uses 1 image, and the image-plus-audio-driven path uses 1 image plus 1 audio clip.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
resolutionselectNo720p-480p, 580p, 720p
aspect_ratioselectNo16:9-16:9, 9:16
enable_prompt_expansionselectNofalse-false, true
seednumberNo-min 0 / max 2147483647-
accelerationselectNonone-none, regular
num_framesnumberNo80min 40 / max 120 / step 4-
frames_per_secondnumberNo16min 4 / max 60 / step 1-
negative_prompttextNo---
num_inference_stepsnumberNo27min 2 / max 40 / step 1-
guidance_scalenumberNo3.5min 1 / max 10 / step 0.1-
shiftnumberNo5min 1 / max 10 / step 0.1-
enable_safety_checkerselectNotrue-true, false

Wan Animate Replace

wan-2-2-animate-replaceAvailable via APIFiles API needed

Video · Choose Wan Animate Replace when the source clip already has the right motion and framing, and only the visible subject should change. Go back to prompt-led generation when the whole shot needs to be reimagined, or use a broader editing project when more than a subject swap is on the table.

Credit pricing

Credit rate: 6–12.5 per second of output The final cost is rounded up to whole credits.

Modes

video-to-video

API input

Files API needed

Inputs

file

Prompt limit

-

Reference media

video, image · 2 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

video: video, image

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
resolutionselectNo480p-480p, 580p, 720p

Wan Animate 2

wan-animate-2This model is coming soon and cannot currently be called through Public API v1.This model is coming soon and cannot currently be called through Public API v1.

Video · Wan Animate 2 is an announced animation entry on Rivya, distinct from existing Wan 2.2 operations. Generation remains unavailable.

Credit pricing

This model is coming soon and cannot currently be called through Public API v1.

Modes

API input

This model is coming soon and cannot currently be called through Public API v1.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model is coming soon and cannot currently be called through Public API v1.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

MiniMax H3

minimax-h3Available via APIText or URL; Files API for reference media

Video · MiniMax H3 is Rivya's 768P or 2K video model for prompt-only generation, first- and last-frame animation, and multimodal reference work. Its reference path can combine images, MP4 or MOV clips, and MP3 or WAV audio, while its credit formula separately accounts for verified input-video time and images beyond the first five.

Credit pricing

The credit cost depends on your inputs and settings. Check the estimate before you start. The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video, reference-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

7000 characters

Reference media

image, video, audio · 15 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "minimax-h3",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "seedance_scene": "text",
    "resolution": "2K",
    "aspect_ratio": "16:9",
    "minimax_single_frame_role": "first",
    "duration": 6
  }
}

Reference rule

Guide frames or multimodal references: Frames accepts 1 or 2 JPEG, PNG, or WebP images up to 30 MB each. Reference accepts up to 9 such images, 3 MP4/MOV videos up to 50 MB and 2-15 seconds each, and 3 MP3/WAV audio files up to 15 MB and 2-15 seconds each; video and audio totals are each capped at 15 seconds, all media is capped at 15 files, and audio cannot be used alone.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
seedance_sceneselectNotext-text, frames, reference
resolutionselectNo2K-768P, 2K
aspect_ratioselectNo16:9-adaptive, 21:9, 16:9, 4:3, 1:1, 3:4, 9:16
minimax_single_frame_roleselectNofirst-first, last
durationnumberNo6min 4 / max 15 / step 1-

Hailuo 2.3

hailuo-2-3Available via APIFiles API needed

Video · MiniMax image-to-video generation from one image and a prompt. Choose Standard or Pro, with 6- or 10-second clips at 768p and 6-second clips at 1080p.

Credit pricing

Credit rate: 25–80 per video The final cost is rounded up to whole credits.

Modes

image-to-video

API input

Files API needed

Inputs

file

Prompt limit

5000 characters

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Reference image: Sign in to upload 1 image for your video. JPG / PNG / WebP · up to 10 MiB (10,485,760 bytes).

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
qualityselectNostandard-standard, pro
resolutionselectNo768P-768P, 1080P
durationselectNo6-6, 10
nsfw_checkerselectNotrue-false, true

Hailuo Pro

hailuo-proAvailable via APIText or URL; Files API for reference media

Video · Create short videos from a written scene or reference images with Hailuo Pro. Rivya uses a fixed 1080P, 6-second configuration for both modes. Add one image for the first frame or two distinct images for the first and last frames. The initial estimate and reservation are 57 credits; see the credit FAQ for final billing.

Credit pricing

Credit rate: 57 per video The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

1500 characters

Reference media

image · 2 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "hailuo-pro",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "prompt_optimizer": "true",
    "nsfw_checker": "true"
  }
}

Reference rule

Reference images: Upload one image for the first frame or two different images for the first and last frames. JPG / PNG / WebP; up to 10 MiB (10,485,760 bytes) per image. Sign in before uploading.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
prompt_optimizerselectNotrue-false, true
nsfw_checkerselectNotrue-false, true

Hailuo Standard

hailuo-standardAvailable via APIText or URL; Files API for reference media

Video · Create 6- or 10-second videos from a text prompt, optionally guided by one or two images. A second image sets an end frame; image-to-video supports 512P, or 768P for 10-second clips.

Credit pricing

Credit rate: 12–50 per video The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

1500 characters

Reference media

image · 2 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "hailuo-standard",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "duration": "6",
    "prompt_optimizer": "true"
  }
}

Reference rule

Reference images: Sign in to upload one or two different images. The first is the start frame; the second, if added, is the end frame. JPG / PNG / WebP · up to 10 MiB (10,485,760 bytes) per image.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
durationselectNo6-6, 10
resolutionselectNo768P-512P, 768P
prompt_optimizerselectNotrue-false, true

Kling 4.0

kling-4-0Available via APIText or URL; Files API for reference media

Video · Choose Keyframe for text or one first-frame image, or Reference for text or up to 10 reference images. Generate 3–30 seconds at 720p / 1080p. Use @first_frame or @image_1, @image_2 to mention supplied images. Reference videos, named subjects, voices, and middle/last frames are unavailable here.

Credit pricing

Credit rate: 28.5–42.5 per second of output The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

8000 characters

Reference media

image · 10 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "kling-4-0",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "kling4_scene": "keyframe",
    "resolution": "720p",
    "duration": 5,
    "aspect_ratio": "16:9"
  }
}

Reference rule

Reference images: Choose Keyframe for text or one first-frame image, or Reference for text or up to 10 reference images. Generate 3–30 seconds at 720p / 1080p. Use @first_frame or @image_1, @image_2 to mention supplied images. Reference videos, named subjects, voices, and middle/last frames are unavailable here.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
kling4_sceneselectNokeyframe-keyframe, reference
resolutionselectNo720p-720p, 1080p
durationnumberNo5min 3 / max 30 / step 1-
aspect_ratioselectNo16:9-16:9, 9:16, 1:1, 21:9

Kling 4.0 Flash

kling-4-0-flashAvailable via APIText or URL; Files API for reference media

Video · Choose Keyframe for text or one first-frame image, or Reference for text or up to 10 reference images. Generate 3–20 seconds at 720p. Use @first_frame or @image_1, @image_2 to mention supplied images. Reference videos, named subjects, voices, and middle/last frames are unavailable here.

Credit pricing

Credit rate: 17 per second of output The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

8000 characters

Reference media

image · 10 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "kling-4-0-flash",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "kling4_scene": "keyframe",
    "resolution": "720p",
    "duration": 5,
    "aspect_ratio": "16:9"
  }
}

Reference rule

Reference images: Choose Keyframe for text or one first-frame image, or Reference for text or up to 10 reference images. Generate 3–20 seconds at 720p. Use @first_frame or @image_1, @image_2 to mention supplied images. Reference videos, named subjects, voices, and middle/last frames are unavailable here.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
kling4_sceneselectNokeyframe-keyframe, reference
resolutionselectNo720p-720p
durationnumberNo5min 3 / max 20 / step 1-
aspect_ratioselectNo16:9-16:9, 9:16, 1:1, 21:9

Kling 3.0

kling-3-0Available via APIText or URL; Files API for reference media

Video · Kuaishou's premium video model for text-to-video and image-to-video, with Standard (720P), Pro (1080P), and 4K tiers, single or multi-shot structure, 3–15s duration, optional audio generation, and up to 2 reference images.

Credit pricing

Credit rate: 15.1–75.5 per second of output The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

2500 characters

Reference media

image · 2 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "kling-3-0",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "mode": "pro",
    "multi_shots": "false",
    "duration": "5",
    "sound": "false",
    "aspect_ratio": "16:9"
  }
}

Reference rule

Reference media: Upload up to 2 image files total for image-to-video projects.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
modeselectNopro-std, pro, 4k
multi_shotsselectNofalse-false, true
durationselectNo5-3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15
soundselectNofalse-false, true
aspect_ratioselectNo16:9-16:9, 9:16, 1:1

Kling 3.0 motion-control

kling-3-0-motion-controlAvailable via APIFiles API needed

Video · Exact 1-image + 1-motion-video project keeps identity and movement roles clear Standard (720P): 22.7 × verified motion-video seconds; Pro (1080P): 30.2 × seconds

Credit pricing

Credit rate: 22.7–30.2 per second of output The final cost is rounded up to whole credits.

Modes

video-to-video

API input

Files API needed

Inputs

file

Prompt limit

2500 characters

Reference media

image, video · 2 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Reference image + motion video: Exact 1-image + 1-motion-video project keeps identity and movement roles clear Standard (720P): 22.7 × verified motion-video seconds; Pro (1080P): 30.2 × seconds

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
modeselectNo720p-720p, 1080p
character_orientationselectNovideo-video, image

Kling 3.0 Turbo

kling-3-turboAvailable via APIText or URL; Files API for reference media

Video · Kling 3.0 Turbo is Rivya's focused Kling video model for prompt-only generation or animation from one still image. Its two paths share 720p and 1080p output with integer durations from 3 to 15 seconds, while aspect-ratio control stays exclusive to text-to-video.

Credit pricing

Credit rate: 20.2–25.2 per second of output The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

2500 characters

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "kling-3-turbo",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "resolution": "720p",
    "duration": 5,
    "aspect_ratio": "16:9"
  }
}

Reference rule

Optional starting image: For image-to-video, upload exactly one JPEG or PNG image up to 10 MB. Text-to-video uses no uploaded media.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
resolutionselectNo720p-720p, 1080p
durationnumberNo5min 3 / max 15 / step 1-
aspect_ratioselectNo16:9-1:1, 9:16, 16:9

Kling O3

kling-o3Available via APIText or URL; Files API for reference media

Video · Verified single-shot generation is available from text, one first frame, or a first-and-last-frame pair.

Credit pricing

Credit rate: 15.1–75.5 per second of output The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

3072 characters

Reference media

image · 2 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "kling-o3",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "resolution": "720p",
    "duration": "5",
    "audio": "false",
    "aspect_ratio": "16:9"
  }
}

Reference rule

Reference media: Upload up to 2 image files total for image-to-video projects.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
resolutionselectNo720p-720p, 1080p, 4k
durationselectNo5-3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15
audioselectNofalse-false, true
aspect_ratioselectNo16:9-16:9, 9:16, 1:1

Kling 2.6

kling-2-6Available via APIText or URL; Files API for reference media

Video · Kuaishou's video model with optional audio generation, 5s/10s clips, and 3 aspect ratios. Strong at human motion and expressive character animation with natural physics.

Credit pricing

Credit rate: 63–252 per video The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

1000 characters

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "kling-2-6",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "duration": "5",
    "sound": "false",
    "aspect_ratio": "1:1"
  }
}

Reference rule

Reference image: Upload up to 1 image file for image projects. JPEG/PNG ≤ 10 MB.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
durationselectNo5-5, 10
soundselectNofalse-false, true
aspect_ratioselectNo1:1-1:1, 16:9, 9:16

Kling 2.6 motion-control

kling-2-6-motion-controlAvailable via APIFiles API needed

Video · Dedicated motion-transfer project for driving one subject from one reference image plus one motion video. Use it when you want a cheaper Kling motion-control pass and can live without the extra scene controls in Kling 3.0 motion-control.

Credit pricing

Credit rate: 12.6–20.2 per second of output The final cost is rounded up to whole credits.

Modes

video-to-video

API input

Files API needed

Inputs

file

Prompt limit

2500 characters

Reference media

image, video · 2 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Reference image + motion video: Upload exactly 1 character image and 1 motion clip. Both should clearly show the head, shoulders, and torso.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
modeselectNo720p-720p, 1080p
character_orientationselectNovideo-video, image

Kling V2.5 Turbo Pro

kling-v2-5-turbo-proAvailable via APIText or URL; Files API for reference media

Video · Kuaishou's Kling V2.5 Turbo Pro video model, now supporting both text-to-video and image-to-video. Public pricing evidence clearly covers both text and image tiers at 5 seconds and 10 seconds, so Rivya maps it directly to 63 / 126 credits.

Credit pricing

Credit rate: 63–126 per video The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

2500 characters

Reference media

image · 2 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "kling-v2-5-turbo-pro",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "duration": "5",
    "aspect_ratio": "16:9",
    "negative_prompt": "Describe what the video should avoid",
    "cfg_scale": 0.5
  }
}

Reference rule

Reference images: Optionally upload up to 2 images to switch into image-to-video mode; the second image maps to the tail frame. JPEG/PNG ≤ 10 MB.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
durationselectNo5-5, 10
aspect_ratioselectNo16:9-16:9, 9:16, 1:1
negative_prompttextNo-max 2500-
cfg_scalenumberNo0.5min 0 / max 1 / step 0.1-

Kling V2.1 Master

kling-v2-1-masterAvailable via APIText or URL; Files API for reference media

Video · Kuaishou's older Kling V2.1 Master video model now supports both text-to-video and image-to-video on Rivya. Current pricing is 252 credits for 5 seconds and 504 credits for 10 seconds.

Credit pricing

Credit rate: 252–504 per video The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "kling-v2-1-master",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "duration": "5",
    "aspect_ratio": "16:9",
    "negative_prompt": "Describe what the video should avoid",
    "cfg_scale": 0.5
  }
}

Reference rule

Reference image: Optionally upload one image to switch into image-to-video mode; leave it empty to stay in text mode. JPEG/PNG ≤ 10 MB.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
durationselectNo5-5, 10
aspect_ratioselectNo16:9-16:9, 9:16, 1:1
negative_prompttextNo-max 500-
cfg_scalenumberNo0.5min 0 / max 1 / step 0.1-

Kling V2.1 Pro

kling-v2-1-proAvailable via APIFiles API needed

Video · Kuaishou's older Kling V2.1 Pro image-to-video model supports a first frame plus an optional tail frame image. Current pricing is 88 credits for 5 seconds and 176 credits for 10 seconds.

Credit pricing

Credit rate: 88–176 per video The final cost is rounded up to whole credits.

Modes

image-to-video

API input

Files API needed

Inputs

file

Prompt limit

5000 characters

Reference media

image · 2 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Reference images: Image mode accepts up to 2 image files, and the second image maps to the tail frame. JPEG/PNG ≤ 10 MB.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
durationselectNo5-5, 10
negative_prompttextNo-max 500-
cfg_scalenumberNo0.5min 0 / max 1 / step 0.1-

Kling V2.1 Standard

kling-v2-1-standardAvailable via APIFiles API needed

Video · Kuaishou's older Kling V2.1 Standard image-to-video model. Current pricing is 51 credits for 5 seconds and 101 credits for 10 seconds.

Credit pricing

Credit rate: 50.5–101 per video The final cost is rounded up to whole credits.

Modes

image-to-video

API input

Files API needed

Inputs

file

Prompt limit

5000 characters

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Reference image: Image mode accepts up to 1 image file as the first frame. JPEG/PNG ≤ 10 MB.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
durationselectNo5-5, 10
negative_prompttextNo-max 500-
cfg_scalenumberNo0.5min 0 / max 1 / step 0.1-

Kling AI Avatar Pro

kling-ai-avatar-proAvailable via APIFiles API needed

Video · Kuaishou's Kling AI Avatar Pro higher-quality talking-avatar model, using one portrait image plus one audio clip to generate lip-synced avatar video. Rivya charges 20.2 credits per second of verified audio, up to 300 seconds, with the total rounded up to a whole credit.

Credit pricing

Credit rate: 20.2 per second of input The final cost is rounded up to whole credits.

Modes

image-to-video

API input

Files API needed

Inputs

file

Prompt limit

5000 characters

Reference media

image, audio · 2 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Portrait image and audio: Upload one portrait image and one audio clip together.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Kling AI Avatar Standard

kling-ai-avatar-standardAvailable via APIFiles API needed

Video · Kuaishou's Kling AI Avatar Standard talking-avatar model, using one portrait image plus one audio clip to generate lip-synced avatar video. Rivya charges 10.1 credits per second of verified audio, up to 300 seconds, with the total rounded up to a whole credit.

Credit pricing

Credit rate: 10.1 per second of input The final cost is rounded up to whole credits.

Modes

image-to-video

API input

Files API needed

Inputs

file

Prompt limit

5000 characters

Reference media

image, audio · 2 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Portrait image and audio: Upload one portrait image and one audio clip together.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Infinitalk

infinitalkAvailable via APIFiles API needed

Video · Infinitalk is a portrait-plus-audio talking-video model. Current pricing is metered by resolution and audio duration: 480p = 3 credits per second and 720p = 12 credits per second.

Credit pricing

Credit rate: 3–12 per second of input The final cost is rounded up to whole credits.

Modes

image-to-video

API input

Files API needed

Inputs

file

Prompt limit

5000 characters

Reference media

image, audio · 2 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Portrait image and audio: Upload one portrait image and one audio clip together.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
resolutionselectNo480p-480p, 720p
seednumberNo-min 10000 / max 1000000-

Runway

runwayAvailable via APIText or URL; Files API for reference media

Video · Runway is a standalone video model that supports both text-to-video and image-to-video. Public pricing evidence currently confirms only 3 generation tiers, so Rivya keeps it on the verified set: `720p_5 = 12`, `720p_10 = 30`, and `1080p_5 = 30`.

Credit pricing

Credit rate: 12–30 per video The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

1800 characters

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "runway",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "duration": "5",
    "quality": "720p",
    "aspectRatio": "16:9",
    "waterMark": "Leave empty to disable watermark"
  }
}

Reference rule

Reference image: Optionally upload one image to switch into image-to-video mode; leave it empty to stay in text mode.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
durationselectNo5-5, 10
qualityselectNo720p-720p, 1080p
aspectRatioselectNo16:9-16:9, 4:3, 1:1, 3:4, 9:16
waterMarktextNo---

Runway Aleph

runway-alephThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Video · Runway Aleph edits an existing video rather than generating a clip from text alone. This route is unavailable on Rivya.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Luma Modify Video

luma-modify-videoThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Video · This exact model version is currently unavailable on Rivya. Choose an available model to continue.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

video · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

video: video

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Grok Imagine Video 1.5 Preview

grok-imagine-videoAvailable via APIText or URL; Files API for reference media

Video · Create a 1–15 second video at 480p, 720p or 1080p from text or one reference image. Prompts support up to 4096 characters.

Credit pricing

Credit rate: 2.4–8 per second of output The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

4096 characters

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "grok-imagine-video",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "duration": 8,
    "resolution": "480p",
    "aspect_ratio": "auto",
    "nsfw_checker": "true"
  }
}

Reference rule

Reference image: Optionally add 1 JPEG, PNG, or WebP image, up to 20 MB.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
durationnumberNo8min 1 / max 15 / step 1-
resolutionselectNo480p-480p, 720p, 1080p
aspect_ratioselectNoauto-auto, 1:1, 16:9, 9:16, 3:2, 2:3, 4:3, 3:4
nsfw_checkerselectNotrue-true, false

Grok Imagine Video Original

grok-imagine-video-originalAvailable via APIText or URL; Files API for reference media

Video · Create a 6–30 second video at 480p, 720p or 1080p from text or reference images. Add up to 7 images at 480p/720p, or 1 at 1080p; each image must be JPEG, PNG or WebP, up to 10 MB. Audio is generated with the video and cannot be disabled. Rivya uses Normal mode.

Credit pricing

Credit rate: 2.4–8 per second of output The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image · 7 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "grok-imagine-video-original",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "resolution": "480p",
    "duration": 6,
    "aspect_ratio": "2:3"
  }
}

Reference rule

Reference images: Create a 6–30 second video at 480p, 720p or 1080p from text or reference images. Add up to 7 images at 480p/720p, or 1 at 1080p; each image must be JPEG, PNG or WebP, up to 10 MB. Audio is generated with the video and cannot be disabled. Rivya uses Normal mode.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
resolutionselectNo480p-480p, 720p, 1080p
durationnumberNo6min 6 / max 30 / step 1-
aspect_ratioselectNo2:3-2:3, 3:2, 1:1, 16:9, 9:16

Sora 2 Pro

sora-2-proThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Video · This exact model version is currently unavailable on Rivya. Choose an available model to continue.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

image: image

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Sora 2

sora-2This service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Video · This exact model version is currently unavailable on Rivya. Choose an available model to continue.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

image: image

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Sora Watermark Remover

sora-watermark-removerThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Video · This exact model version is currently unavailable on Rivya. Choose an available model to continue.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Topaz Video Upscaler

topaz-video-upscaleAvailable via APIFiles API needed

Video · One uploaded video, no prompt, with a 1x, 2x, or 4x upscale factor; 2x is the default.

Credit pricing

Credit rate: 8–14 per second of input The final cost is rounded up to whole credits.

Modes

video-to-video

API input

Files API needed

Inputs

file

Prompt limit

-

Reference media

video · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Source video: Rivya accepts one MP4, MOV, or MKV file up to 10MB; the upstream interface allows 50MB, so Rivya uses the stricter limit.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
upscale_factorselectNo2-1, 2, 4

Veo3.1 Quality

veo3-1-qualityAvailable via APIText or URL; Files API for reference media

Video · Generate a clip from text or uploaded frames, then review the result before sharing it.

Credit pricing

Credit rate: 225–285 per video The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image · 2 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "veo3-1-quality",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "generationType": "FIRST_AND_LAST_FRAMES_2_VIDEO",
    "aspectRatio": "16:9",
    "resolution": "720p",
    "duration": "8",
    "watermark": "Optional watermark text",
    "enableTranslation": "true"
  }
}

Reference rule

Start and end frames: Upload no more than 2 JPEG or PNG images, up to 10 MiB each. The first image is the start frame; a second image sets the end frame.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
generationTypeselectNoFIRST_AND_LAST_FRAMES_2_VIDEO-FIRST_AND_LAST_FRAMES_2_VIDEO
aspectRatioselectNo16:9-16:9, 9:16, Auto
resolutionselectNo720p-720p, 1080p, 4k
durationselectNo8-4, 6, 8
watermarktextNo---
enableTranslationselectNotrue-true, false

Veo3.1 Fast

veo3-1-fastAvailable via APIText or URL; Files API for reference media

Video · Generate video from a prompt, start and end frames, or reference images. Check the picture and sound before using a result.

Credit pricing

Credit rate: 30–90 per video The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video, reference-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image · 3 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "veo3-1-fast",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "generationType": "TEXT_2_VIDEO",
    "aspectRatio": "16:9",
    "resolution": "720p",
    "duration": "8",
    "watermark": "Optional watermark text",
    "enableTranslation": "true"
  }
}

Reference rule

Reference images: Upload JPEG or PNG images up to 10 MiB each. Use a starting frame or a start and end pair, or choose reference image mode with up to 3 images. Reference image mode requires 8 seconds at 16:9 or 9:16.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
generationTypeselectNoTEXT_2_VIDEO-TEXT_2_VIDEO, FIRST_AND_LAST_FRAMES_2_VIDEO, REFERENCE_2_VIDEO
aspectRatioselectNo16:9-16:9, 9:16, Auto
resolutionselectNo720p-720p, 1080p, 4k
durationselectNo8-4, 6, 8
watermarktextNo---
enableTranslationselectNotrue-true, false

Veo3.1 Lite

veo3-1-liteAvailable via APIText or URL; Files API for reference media

Video · Make a video draft from a prompt, start and end frames, or reference images. Review the motion and any sound before using it.

Credit pricing

Credit rate: 15–75 per video The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video, reference-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image · 3 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "veo3-1-lite",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "generationType": "FIRST_AND_LAST_FRAMES_2_VIDEO",
    "aspectRatio": "16:9",
    "resolution": "720p",
    "duration": "8",
    "watermark": "Optional watermark text",
    "enableTranslation": "true"
  }
}

Reference rule

Frame or reference images: Use JPEG or PNG files up to 10 MiB each. Frame mode uses a start image and an optional end image. Reference mode accepts up to 3 images and requires 8 seconds with an aspect ratio of 16:9 or 9:16.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
generationTypeselectNoFIRST_AND_LAST_FRAMES_2_VIDEO-FIRST_AND_LAST_FRAMES_2_VIDEO, REFERENCE_2_VIDEO
aspectRatioselectNo16:9-16:9, 9:16, Auto
resolutionselectNo720p-720p, 1080p, 4k
durationselectNo8-4, 6, 8
watermarktextNo---
enableTranslationselectNotrue-true, false

Vidu Q4

vidu-q4Available via APIFiles API needed

Video · Vidu Q4 — Generate a video from one starting image, with a prompt describing the motion and scene.

Credit pricing

Credit rate: 7.2–62.4 per second of output The final cost is rounded up to whole credits.

Modes

image-to-video

API input

Files API needed

Inputs

file

Prompt limit

20000 characters

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Reference image: Upload one JPEG, PNG, or WebP starting image, no larger than 50 MB.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
resolutionselectNo720p-540p, 720p, 1080p, 2k, 4k
durationnumberNo5min 3 / max 16 / step 1-
audioselectNotrue-true, false

Gemini Omni 1.2 Flash

gemini-omni-flash-1-2Available via APIText or URL; Files API for reference media

Video · Create video from text, references or frames, or edit one source video with Gemini Omni 1.2 Flash.

Credit pricing

Credit rate: 4–22 per second of output The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video, video-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

20000 characters

Reference media

image, video · 7 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "gemini-omni-flash-1-2",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "omni_scene": "reference",
    "resolution": "720p",
    "duration": 8,
    "aspect_ratio": "16:9",
    "seeds": 10000
  }
}

Reference rule

Reference files: Choose 360p, 720p, 1080p or 4k. Text, references and one first frame allow 3–15 seconds; a first/last-frame pair allows 3–10. Video edit follows one uploaded video of up to 15 seconds and sends no duration control.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
omni_sceneselectNoreference-reference, frames, edit
resolutionselectNo720p-360p, 720p, 1080p, 4k
durationnumberNo8min 3 / max 15 / step 1-
aspect_ratioselectNo16:9-16:9, 9:16, auto
seedsnumberNo-min 10000 / max 99999 / step 1-

Gemini Omni 1.1 Flash

gemini-omni-flash-1-1Available via APIText or URL; Files API for reference media

Video · Create video from text, reference images, or a source video. Choose 16:9 or 9:16 and 4, 6, 8, or 10 seconds; with video input, the model determines output length.

Credit pricing

Credit rate: 45–180 per video The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video, video-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

20000 characters

Reference media

image, video · 7 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "gemini-omni-flash-1-1",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "reference_mode": "reference",
    "resolution": "720p",
    "duration": "4",
    "aspect_ratio": "16:9",
    "seed": 0,
    "video_start": 0,
    "video_ends": 0
  }
}

Reference rule

Reference assets: Use up to seven images at 20 MB each, or one video up to 95 MB and 30 seconds with up to five images. Select a video segment no longer than 10 seconds. Audio and character asset IDs are not supported. Frames mode accepts one start image or a start/end pair, without other references.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
reference_modeselectNoreference-reference, frames
resolutionselectNo720p-360p, 720p, 1080p, 4k
durationselectNo4-4, 6, 8, 10
aspect_ratioselectNo16:9-16:9, 9:16
seednumberNo-min 0 / max 2147483647-
video_startnumberNo-min 0 / max 30-
video_endsnumberNo-min 0 / max 30-

Gemini Omni Video

gemini-omni-videoAvailable via APIText or URL; Files API for reference media

Video · Create video from text, reference images, or a source video. Choose 16:9 or 9:16 and 4, 6, 8, or 10 seconds; with video input, the model determines output length.

Credit pricing

Credit rate: 45–180 per video The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video, video-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

20000 characters

Reference media

image, video · 7 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "gemini-omni-video",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "resolution": "720p",
    "duration": "4",
    "aspect_ratio": "16:9",
    "seed": 0,
    "video_start": 0,
    "video_ends": 0
  }
}

Reference rule

Reference assets: Use up to seven images at 20 MB each, or one video up to 95 MB and 30 seconds with up to five images. Select a video segment no longer than 10 seconds. Audio and character asset IDs are not supported.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
resolutionselectNo720p-720p, 1080p, 4k
durationselectNo4-4, 6, 8, 10
aspect_ratioselectNo16:9-16:9, 9:16
seednumberNo-min 0 / max 2147483647-
video_startnumberNo-min 0 / max 30-
video_endsnumberNo-min 0 / max 30-

OmniHuman 2.0

omnihuman-2-0Available via APIFiles API needed

Video · Create an audio-driven video from one image and 4–300 seconds of audio. An optional prompt supports up to 300 characters.

Credit pricing

Credit rate: 14 per second of input The final cost is rounded up to whole credits.

Modes

image-audio-to-video

API input

Files API needed

Inputs

file

Prompt limit

300 characters

Reference media

image, audio · 2 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Image and audio: Upload 1 JPEG, PNG, or WebP image under 30 MB and 1 MP3 or WAV audio file under 15 MB. Audio must be 4–300 seconds.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
enable_watermarkselectNotrue-true, false

OmniHuman 1.5

omnihuman-1-5Available via APIFiles API needed

Video · OmniHuman 1.5 turns exactly one portrait image and one audio clip into an audio-driven character video at 720p or 1080p, with credits based on verified audio duration.

Credit pricing

Credit rate: 27 per second of input The final cost is rounded up to whole credits.

Modes

image-audio-to-video

API input

Files API needed

Inputs

file

Prompt limit

300 characters

Reference media

image, audio · 2 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Portrait image and audio: Upload exactly one JPEG, PNG, or WebP portrait up to 10 MB and one MP3, M4A, WAV, AAC, or OGG audio file up to 10 MB and under 60 seconds.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
output_resolutionselectNo1080-720, 1080
pe_fast_modeselectNofalse-false, true
seednumberNo-1--

Volcengine Video Lip Sync

volcengine-video-lip-syncAvailable via APIFiles API needed

Video · Volcengine Video Lip Sync replaces the spoken performance in one existing video with one target vocal track, using Lite or Basic processing and duration-based credits. The output is MP4 at 25 FPS, and its duration follows the target audio.

Credit pricing

Credit rate: 8 per second of input The final cost is rounded up to whole credits.

Modes

video-audio-to-video

API input

Files API needed

Inputs

file

Prompt limit

-

Reference media

video, audio · 2 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Source video and target audio: Upload exactly one MP4 or MOV video up to 95 MB and one MP3, M4A, WAV, AAC, or OGG vocal track up to 10 MB.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
modeselectYeslite-lite, basic
separate_vocalselectNofalse-false, true
open_scenedetselectNofalse-false, true
align_audioselectNotrue-true, false
align_audio_reverseselectNofalse-false, true
templ_start_secondsnumberNo0min 0-

PixVerse V6

pixverse-v6Available via APIText or URL; Files API for reference media

Video · Create a 1–15 second video from text, 1–2 opening/ending images, or up to 7 reference images. Choose 360p, 540p, 720p or 1080p and optional generated audio.

Credit pricing

Credit rate: 4–18.4 per second of output The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video, reference-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image · 7 files

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "pixverse-v6",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "pixverse_scene": "text",
    "quality": "720p",
    "aspect_ratio": "16:9",
    "duration": 5,
    "generate_audio_switch": "false",
    "generate_multi_clip_switch": "false",
    "seed": 0
  }
}

Reference rule

Reference images: Create a 1–15 second video from text, 1–2 opening/ending images, or up to 7 reference images. Choose 360p, 540p, 720p or 1080p and optional generated audio.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
pixverse_sceneselectNotext-text, image, reference
qualityselectNo720p-360p, 540p, 720p, 1080p
aspect_ratioselectNo16:9-16:9, 4:3, 1:1, 3:4, 9:16, 2:3, 3:2, 21:9
durationnumberNo5min 1 / max 15 / step 1-
generate_audio_switchselectNofalse-false, true
generate_multi_clip_switchselectNofalse-false, true
seednumberNo-min 0 / max 2147483647 / step 1-

PixVerse V6 Max

pixverse-v6-maxAvailable via APIText or URL; Files API for reference media

Video · Create a 1–15 second video from text or one image. Choose 360p, 540p, 720p or 1080p, with optional generated audio.

Credit pricing

Credit rate: 4.8–22.4 per second of output The final cost is rounded up to whole credits.

Modes

text-to-video, image-to-video

API input

Text or URL; Files API for reference media

Inputs

text, file

Prompt limit

5000 characters

Reference media

image · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "pixverse-v6-max",
  "prompt": "A camera slowly circles a product bottle on a studio table.",
  "params": {
    "quality": "720p",
    "duration": 5,
    "generate_audio_switch": "false",
    "aspect_ratio": "16:9",
    "seed": 0
  }
}

Reference rule

First-frame image: Optionally upload 1 JPEG, PNG, or WebP image, up to 20 MB and 10000 pixels per side.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Text or URL modes can run without reference uploads. For modes that use reference media, upload the files through Files API first. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
qualityselectNo720p-360p, 540p, 720p, 1080p
durationnumberNo5min 1 / max 15 / step 1-
generate_audio_switchselectNofalse-false, true
aspect_ratioselectNo16:9-16:9, 4:3, 1:1, 3:4, 9:16, 2:3, 3:2, 21:9
seednumberNo-min 0 / max 2147483647 / step 1-

Rivya Video Upscale

video-upscaleAvailable via APIFiles API needed

Video · Enhance one video within a 1080P, 2K or 4K output box while keeping its aspect ratio.

Credit pricing

Credit rate: 1.5–6 per second of input The final cost is rounded up to whole credits.

Modes

video-to-video

API input

Files API needed

Inputs

file

Prompt limit

-

Reference media

video · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Reference files: Upload MP4, MOV or WebM up to 50 MB and 60 seconds. The output fits the chosen box without cropping; larger sources may be reduced. The operation preserves the source audio, frame rate and frame count.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

keyTypeRequiredDefaultRangeOptions
resolutionselectNo2K-1080P, 2K, 4K

Suno Music

suno-musicAvailable via APIText input ready

Audio · Create music in Custom mode. A title (up to 80 characters) and style (up to 1000) are required. For vocals, enter lyrics up to 5000 characters; for instrumental music, leave lyrics empty. Choose V6, V6 Mini or V6 Wild and a target duration of 10–360 seconds.

Credit pricing

Credit rate: 12 per request The final cost is rounded up to whole credits.

Modes

text-to-audio

API input

Text input ready

Inputs

text

Prompt limit

5000 characters

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "suno-music",
  "prompt": "Welcome. Here is a quick introduction to the product.",
  "params": {
    "model": "V6",
    "instrumental": "false",
    "duration": 30,
    "title": "Suno Music",
    "style": "Acoustic folk, gentle rhythm, warm vocals"
  }
}

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Parameters

keyTypeRequiredDefaultRangeOptions
modelselectNoV6-V6, V6_MINI, V6_WILD
instrumentalselectNotrue-true, false
titletextYes---
styletextYes---
durationnumberNo30min 10 / max 360 / step 1-

Suno Sounds

suno-soundsAvailable via APIText input ready

Audio · Create ambience, loops and sound textures from a text description.

Credit pricing

Credit rate: 2.5 per request The final cost is rounded up to whole credits.

Modes

text-to-audio

API input

Text input ready

Inputs

text

Prompt limit

500 characters

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "suno-sounds",
  "prompt": "Welcome. Here is a quick introduction to the product.",
  "params": {
    "model": "V6",
    "soundLoop": "false",
    "soundTempo": 1,
    "soundKey": "Any"
  }
}

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Parameters

keyTypeRequiredDefaultRangeOptions
modelselectNoV6-V6, V6_MINI, V6_WILD
soundLoopselectNofalse-false, true
soundTemponumberNo-min 1 / max 300 / step 1-
soundKeyselectNoAny-Any, Cm, C#m, Dm, D#m, Em, Fm, F#m, Gm, G#m, Am, A#m, Bm, C, C#, D, D#, E, F, F#, G, G#, A, A#, B

Suno Lyrics

suno-lyricsAvailable via APIText input ready

Audio · Draft song lyrics from a theme, mood or story with Suno Lyrics.

Credit pricing

Credit rate: 0.4 per request The final cost is rounded up to whole credits.

Modes

text-to-audio

API input

Text input ready

Inputs

text

Prompt limit

200 characters

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "suno-lyrics",
  "prompt": "Welcome. Here is a quick introduction to the product."
}

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Suno Speech

suno-speechAvailable via APIText input ready

Audio · Suno Speech creates spoken audio with tone, rhythm and emotion. Simple mode writes a script from your scene; Advanced mode reads the script you provide.

Credit pricing

Credit rate: 12 per request The final cost is rounded up to whole credits.

Modes

text-to-speech

API input

Text input ready

Inputs

text

Prompt limit

5000 characters

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "suno-speech",
  "prompt": "Welcome. Here is a quick introduction to the product.",
  "params": {
    "customMode": "false"
  }
}

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Parameters

keyTypeRequiredDefaultRangeOptions
customModeselectNofalse-false, true
tonetextNo-max 1000-
varietynumberNo-min 0 / max 4 / step 1-
vocalGenderselectNof-f, m
backgroundMusicselectNofalse-false, true

ElevenLabs Dialogue V3

elevenlabs-dialogue-v3Available via APIText input ready

Audio · Generate audio from a dialogue script on Rivya, with a preset voice selected for each line. Set Stability and an optional language code for the request.

Credit pricing

Credit rate: 14 per 1,000 characters The final cost is rounded up to whole credits.

Modes

dialogue-to-audio

API input

Text input ready

Inputs

text

Prompt limit

5000 characters

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "elevenlabs-dialogue-v3",
  "prompt": "Welcome. Here is a quick introduction to the product.",
  "params": {
    "default_voice": "Adam",
    "stability": "0.5",
    "language_code": "Optional language code",
    "dialogue": [
      {
        "voice": "Adam",
        "text": "Welcome to our product demo."
      },
      {
        "voice": "Alice",
        "text": "Let's look at how it works."
      }
    ]
  }
}

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Parameters

keyTypeRequiredDefaultRangeOptions
default_voiceselectNoAdam-Adam, Alice, Bill, Brian, Callum, Charlie, Chris, Daniel, Eric, George, Harry, Jessica, Laura, Liam, Lily, Matilda, River, Roger, Sarah, Will
stabilityselectNo0.5-0, 0.5, 1
language_codetextNo---

ElevenLabs Turbo 2.5

elevenlabs-turbo-2-5Available via APIText input ready

Audio · Turn written scripts into spoken audio with ElevenLabs Turbo 2.5 on Rivya. Choose a preset voice and adjust delivery for demos, walkthroughs, and narration.

Credit pricing

Credit rate: 6 per 1,000 characters The final cost is rounded up to whole credits.

Modes

text-to-speech

API input

Text input ready

Inputs

text

Prompt limit

5000 characters

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "elevenlabs-turbo-2-5",
  "prompt": "Welcome. Here is a quick introduction to the product.",
  "params": {
    "voice": "TX3LPaxmHKxFdv7VOQHJ",
    "stability": 0.5,
    "similarity_boost": 0.75,
    "style": 0,
    "speed": 1,
    "timestamps": "false",
    "previous_text": "Text before this passage (optional)",
    "next_text": "Text after this passage (optional)",
    "language_code": "Optional language code"
  }
}

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Parameters

keyTypeRequiredDefaultRangeOptions
voiceselectYesTX3LPaxmHKxFdv7VOQHJ-TX3LPaxmHKxFdv7VOQHJ, cgSgspJ2msm6clMCkdW9, pNInz6obpgDQGcFmaJgB, BIvP0GN1cAtSRTxNHnWS, aMSt68OGf4xUZAnLpTU8
stabilitynumberNo0.5min 0 / max 1 / step 0.05-
similarity_boostnumberNo0.75min 0 / max 1 / step 0.05-
stylenumberNo0min 0 / max 1 / step 0.05-
speednumberNo1min 0.7 / max 1.2 / step 0.05-
timestampsselectNofalse-false, true
previous_texttextNo---
next_texttextNo---
language_codetextNo---

ElevenLabs Multilingual V2

elevenlabs-multilingual-v2Available via APIText input ready

Audio · Generate speech from written scripts with ElevenLabs Multilingual V2. Choose a voice preset and adjust delivery for localized demos, lessons, and narration.

Credit pricing

Credit rate: 12 per 1,000 characters The final cost is rounded up to whole credits.

Modes

text-to-speech

API input

Text input ready

Inputs

text

Prompt limit

5000 characters

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "elevenlabs-multilingual-v2",
  "prompt": "Welcome. Here is a quick introduction to the product.",
  "params": {
    "voice": "TX3LPaxmHKxFdv7VOQHJ",
    "stability": 0.5,
    "similarity_boost": 0.75,
    "style": 0,
    "speed": 1,
    "timestamps": "false",
    "previous_text": "Text before this passage",
    "next_text": "Text after this passage"
  }
}

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Parameters

keyTypeRequiredDefaultRangeOptions
voiceselectYesTX3LPaxmHKxFdv7VOQHJ-TX3LPaxmHKxFdv7VOQHJ, cgSgspJ2msm6clMCkdW9, pNInz6obpgDQGcFmaJgB, BIvP0GN1cAtSRTxNHnWS, aMSt68OGf4xUZAnLpTU8
stabilitynumberNo0.5min 0 / max 1 / step 0.05-
similarity_boostnumberNo0.75min 0 / max 1 / step 0.05-
stylenumberNo0min 0 / max 1 / step 0.05-
speednumberNo1min 0.7 / max 1.2 / step 0.05-
timestampsselectNofalse-false, true
previous_texttextNo---
next_texttextNo---

ElevenLabs Sound Effect V2

elevenlabs-sound-effect-v2This service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Audio · ElevenLabs Sound Effects V2 generates sounds from a written description. This version is paused on Rivya.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

ElevenLabs Audio Isolation

elevenlabs-audio-isolationThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Audio · Audio Isolation separates the voice in an existing recording from background sound. Rivya’s route is unavailable.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

audio · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

audio: audio

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Gemini 3.8 Flash TTS

gemini-3-8-flash-ttsAvailable via APIText input ready

Audio · Turn a short script into speech with Gemini 3.8 Flash TTS. Choose one preset voice for a single-speaker recording.

Credit pricing

Base rate per million tokens: 70 credits for input, 1,260 for output The final cost is rounded up to whole credits.

Modes

text-to-speech

API input

Text input ready

Inputs

text

Prompt limit

Up to 1,024 UTF-8 bytes

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "gemini-3-8-flash-tts",
  "prompt": "Welcome. Here is a quick introduction to the product.",
  "params": {
    "voice_name": "Zephyr"
  }
}

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Parameters

keyTypeRequiredDefaultRangeOptions
voice_nameselectNoZephyr-Zephyr, Puck, Kore

Gemini 3.8 Flash-Lite TTS

gemini-3-8-flash-lite-ttsAvailable via APIText input ready

Audio · Turn a short script into speech with Gemini 3.8 Flash-Lite TTS. Choose one preset voice for a single-speaker recording.

Credit pricing

Base rate per million tokens: 70 credits for input, 840 for output The final cost is rounded up to whole credits.

Modes

text-to-speech

API input

Text input ready

Inputs

text

Prompt limit

Up to 1,024 UTF-8 bytes

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "gemini-3-8-flash-lite-tts",
  "prompt": "Welcome. Here is a quick introduction to the product.",
  "params": {
    "voice_name": "Zephyr"
  }
}

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Parameters

keyTypeRequiredDefaultRangeOptions
voice_nameselectNoZephyr-Zephyr, Puck, Kore

Gemini 3.1 Flash TTS

gemini-3-1-flash-ttsAvailable via APIText input ready

Audio · Turn a short script into speech with Gemini 3.1 Flash TTS. Choose one preset voice for a single-speaker recording.

Credit pricing

Base rate per million tokens: 140 credits for input, 2,800 for output The final cost is rounded up to whole credits.

Modes

text-to-speech

API input

Text input ready

Inputs

text

Prompt limit

Up to 1,024 UTF-8 bytes

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "gemini-3-1-flash-tts",
  "prompt": "Welcome. Here is a quick introduction to the product.",
  "params": {
    "voice_name": "Zephyr"
  }
}

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Parameters

keyTypeRequiredDefaultRangeOptions
voice_nameselectNoZephyr-Zephyr, Puck, Kore

Gemini 2.5 Pro TTS

gemini-2-5-pro-ttsAvailable via APIText input ready

Audio · Turn a short script into speech with Gemini 2.5 Pro TTS. Choose one preset voice for a single-speaker recording.

Credit pricing

Base rate per million tokens: 140 credits for input, 2,800 for output The final cost is rounded up to whole credits.

Modes

text-to-speech

API input

Text input ready

Inputs

text

Prompt limit

Up to 1,024 UTF-8 bytes

Reference media

-

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

{
  "model": "gemini-2-5-pro-tts",
  "prompt": "Welcome. Here is a quick introduction to the product.",
  "params": {
    "voice_name": "Zephyr"
  }
}

Reference rule

-

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

Parameters

keyTypeRequiredDefaultRangeOptions
voice_nameselectNoZephyr-Zephyr, Puck, Kore

Nemotron 3.5 Transcribe

nemotron-3-5-transcribeAvailable via APIFiles API needed

Audio · Transcribe one MP3 or WAV file up to 60 minutes and 95 MB. Rivya returns the transcript as text; word timestamps, translation and video input are unavailable here.

Credit pricing

Credit rate: 0.005 per second of input The final cost is rounded up to whole credits.

Modes

audio-to-text

API input

Files API needed

Inputs

file

Prompt limit

-

Reference media

audio · 1 file

Endpoint

POST /api/v1/generations

Authentication

Bearer API key

Idempotency-Key

Recommended

Request body

No runnable Public API example is shown for this model.

Reference rule

Audio: Transcribe one MP3 or WAV file up to 60 minutes and 95 MB. Rivya returns the transcript as text; word timestamps, translation and video input are unavailable here.

Result

The create response includes a task ID. Poll GET /api/v1/generations/{taskId} until the status is succeeded or failed.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, not_found

This model needs reference files. Upload them through /api/v1/files and follow the model's API guide to build your request. Files API

Parameters

No additional model controls are listed. Check reference-input requirements separately.

GPT 6.1 Sol

gpt-6-1-sol-chatAvailable via APIText input ready

Chat · Chat using text, without attachments or web search. Credits are reserved before sending using a conservative output allowance; final billing uses confirmed consumption.

Credit pricing

Base rate per million tokens: 120 credits for input, 600 for output The final cost is rounded up to whole credits.

Modes

responses

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "gpt-6-1-sol-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

GPT 6.1 Terra

gpt-6-1-terra-chatAvailable via APIText input ready

Chat · Chat using text, without attachments or web search. Credits are reserved before sending using a conservative output allowance; final billing uses confirmed consumption.

Credit pricing

Base rate per million tokens: 60 credits for input, 300 for output The final cost is rounded up to whole credits.

Modes

responses

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "gpt-6-1-terra-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

GPT 6 Astra

gpt-6-astra-chatAvailable via APIText input ready

Chat · Chat using text, without attachments or web search. Credits are reserved before sending using a conservative output allowance; final billing uses confirmed consumption.

Credit pricing

Base rate per million tokens: 560 credits for input, 2,800 for output The final cost is rounded up to whole credits.

Modes

responses

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "gpt-6-astra-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

GPT 6 Sol

gpt-6-sol-chatAvailable via APIText input ready

Chat · Chat using text, without attachments or web search. Credits are reserved before sending using a conservative output allowance; final billing uses confirmed consumption.

Credit pricing

Base rate per million tokens: 120 credits for input, 600 for output The final cost is rounded up to whole credits.

Modes

responses

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "gpt-6-sol-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

GPT 6 Luna

gpt-6-luna-chatAvailable via APIText input ready

Chat · Chat using text, without attachments or web search. Credits are reserved before sending using a conservative output allowance; final billing uses confirmed consumption.

Credit pricing

Base rate per million tokens: 6 credits for input, 30 for output The final cost is rounded up to whole credits.

Modes

responses

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "gpt-6-luna-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

GPT-5.6 Luna

gpt-5-6-luna-chatAvailable via APIText input ready

Chat · Chat using text, without attachments or web search. Credits are reserved before sending using a conservative output allowance; final billing uses confirmed consumption.

Credit pricing

Base rate per million tokens: 11.2 credits for input, 67.2 for output The final cost is rounded up to whole credits.

Modes

responses

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "gpt-5-6-luna-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

GPT-5.6 Terra

gpt-5-6-terra-chatAvailable via APIText input ready

Chat · Chat using text, without attachments or web search. Credits are reserved before sending using a conservative output allowance; final billing uses confirmed consumption.

Credit pricing

Base rate per million tokens: 112 credits for input, 672 for output The final cost is rounded up to whole credits.

Modes

responses

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "gpt-5-6-terra-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

GPT-5.6 Sol

gpt-5-6-sol-chatAvailable via APIText input ready

Chat · Chat using text, without attachments or web search. Credits are reserved before sending using a conservative output allowance; final billing uses confirmed consumption.

Credit pricing

Base rate per million tokens: 280 credits for input, 1,680 for output The final cost is rounded up to whole credits.

Modes

responses

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "gpt-5-6-sol-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

GPT-5.5

gpt-5-5-chatAvailable via APIText input ready

Chat · Chat using text, without attachments or web search. Credits are reserved before sending using a conservative output allowance; final billing uses confirmed consumption.

Credit pricing

Base rate per million tokens: 140 credits for input, 840 for output The final cost is rounded up to whole credits.

Modes

responses

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "gpt-5-5-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

GPT-5.4

gpt-5-4-chatThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Chat · This exact model version is currently unavailable on Rivya. Choose an available model to continue.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

GPT-5.4 Codex

gpt-5.4-codex-chatThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Chat · This exact model version is currently unavailable on Rivya. Choose an available model to continue.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

GPT-5.3 Codex

gpt-5-3-codex-chatThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Chat · This exact model version is currently unavailable on Rivya. Choose an available model to continue.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

GPT-5.2

gpt-5-2-chatAvailable via APIText input ready

Chat · Chat using text, without attachments or web search. Credits are reserved before sending using a conservative output allowance; final billing uses confirmed consumption. The answer appears once the complete response is ready.

Credit pricing

Base rate per million tokens: 87.5 credits for input, 700 for output The final cost is rounded up to whole credits.

Modes

chat-completion

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "gpt-5-2-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

GPT-5.2 Codex

gpt-5-2-codex-chatThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Chat · This exact model version is currently unavailable on Rivya. Choose an available model to continue.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

GPT-5.1 Codex

gpt-5-1-codex-chatThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Chat · This exact model version is currently unavailable on Rivya. Choose an available model to continue.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

GPT-5 Codex

gpt-5-codex-chatThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Chat · This exact model version is currently unavailable on Rivya. Choose an available model to continue.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Gemini 4 Argon

gemini-4-argon-chatAvailable via APIText input ready

Chat · Chat using text, without attachments or web search. Credits are reserved before sending using a conservative output allowance; final billing uses confirmed consumption.

Credit pricing

Base rate per million tokens: 160 credits for input, 800 for output The final cost is rounded up to whole credits.

Modes

chat-completion

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "gemini-4-argon-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Gemini 4 Carbon

gemini-4-carbon-chatAvailable via APIText input ready

Chat · Use Gemini 4 Carbon for a text-only conversation with a single complete response.

Credit pricing

Base rate per million tokens: 160 credits for input, 800 for output The final cost is rounded up to whole credits.

Modes

chat-completion

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "gemini-4-carbon-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Gemini 3.8 Flash

gemini-3-8-flash-chatAvailable via APIText input ready

Chat · Chat using text, without attachments or web search. Credits are reserved before sending using a conservative output allowance; final billing uses confirmed consumption.

Credit pricing

Base rate per million tokens: 45 credits for input, 225 for output The final cost is rounded up to whole credits.

Modes

chat-completion

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "gemini-3-8-flash-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Gemini 3.7 Flash

gemini-3-7-flash-chatThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Chat · Gemini 3.7 Flash now maps to Gemini 3.8 Flash as a compatibility alias. Rivya keeps this entry as a reference rather than a separate callable model. Choose Gemini 3.8 Flash to start a new conversation.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Gemini 3.6 Flash

gemini-3-6-flash-chatAvailable via APIText input ready

Chat · Chat using text, without attachments or web search. Credits are reserved before sending using a conservative output allowance; final billing uses confirmed consumption.

Credit pricing

Base rate per million tokens: 45 credits for input, 225 for output The final cost is rounded up to whole credits.

Modes

chat-completion

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "gemini-3-6-flash-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Gemini 3.5 Flash

gemini-3-5-flash-chatThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Chat · Gemini 3.5 Flash Chat is unavailable on Rivya.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Gemini 3.1 Pro

gemini-3-1-pro-chatAvailable via APIText input ready

Chat · Chat using text, without attachments or web search. Credits are reserved before sending using a conservative output allowance; final billing uses confirmed consumption.

Credit pricing

Base rate per million tokens: 100 credits for input, 700 for output The final cost is rounded up to whole credits.

Modes

chat-completion

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "gemini-3-1-pro-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Gemini 3 Pro

gemini-3-pro-chatAvailable via APIText input ready

Chat · Chat using text, without attachments or web search. Credits are reserved before sending using a conservative output allowance; final billing uses confirmed consumption.

Credit pricing

Base rate per million tokens: 50 credits for input, 350 for output The final cost is rounded up to whole credits.

Modes

chat-completion

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "gemini-3-pro-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Gemini 3 Flash

gemini-3-flash-chatThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Chat · This exact model version is currently unavailable on Rivya. Choose an available model to continue.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Gemini 2.5 Pro

gemini-2-5-pro-chatThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Chat · Gemini 2.5 Pro Chat is unavailable on Rivya; Gemini 2.5 Pro TTS is a separate speech model.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Gemini 2.5 Flash

gemini-2-5-flash-chatThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Chat · This exact model version is currently unavailable on Rivya. Choose an available model to continue.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Claude Opus 5.5

claude-opus-5-5-chatAvailable via APIText input ready

Chat · Send text messages and continue the conversation. Output is limited to 4096 tokens per reply.

Credit pricing

Base rate per million tokens: 320 credits for input, 1,600 for output The final cost is rounded up to whole credits.

Modes

claude-messages

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "claude-opus-5-5-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Claude Sonnet 5.5

claude-sonnet-5-5-chatAvailable via APIText input ready

Chat · Send text messages and continue the conversation. Output is limited to 4096 tokens per reply.

Credit pricing

Base rate per million tokens: 160 credits for input, 800 for output The final cost is rounded up to whole credits.

Modes

claude-messages

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "claude-sonnet-5-5-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Claude Haiku 5.5

claude-haiku-5-5-chatAvailable via APIText input ready

Chat · Claude Haiku 5.5 — Use text conversations to organize ideas, draft content, or discuss questions.

Credit pricing

Base rate per million tokens: 12 credits for input, 60 for output The final cost is rounded up to whole credits.

Modes

claude-messages

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "claude-haiku-5-5-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Claude Fable 5.2

claude-fable-5-2-chatThis model is coming soon and cannot currently be called through Public API v1.This model is coming soon and cannot currently be called through Public API v1.

Chat · Claude Fable 5.2 is an announced chat entry on Rivya. This page helps distinguish the preview from older Fable names and current Claude choices.

Credit pricing

This model is coming soon and cannot currently be called through Public API v1.

Modes

API input

This model is coming soon and cannot currently be called through Public API v1.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

This model is coming soon and cannot currently be called through Public API v1.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Claude Fable 5.1

claude-fable-5-1-chatThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Chat · This model is listed but cannot be used on Rivya while its current generation contract and credit price remain unconfirmed.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Claude Fable 5

claude-fable-5-chatAvailable via APIText input ready

Chat · Text chat on Rivya for drafting from notes, revising copy, and adapting a message for a different audience.

Credit pricing

Base rate per million tokens: 800 credits for input, 4,000 for output The final cost is rounded up to whole credits.

Modes

claude-messages

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "claude-fable-5-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Claude Sonnet 5

claude-sonnet-5-chatAvailable via APIText input ready

Chat · Use Claude Sonnet 5 on Rivya to discuss pasted requirements, draft implementation briefs, and prepare technical handoffs.

Credit pricing

Base rate per million tokens: 170 credits for input, 855 for output The final cost is rounded up to whole credits.

Modes

claude-messages

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "claude-sonnet-5-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Claude Opus 5

claude-opus-5-chatAvailable via APIText input ready

Chat · A text-only Claude model for online chat on Rivya. Paste a draft or a pair of proposals and ask for a rewrite, comparison, or list of questions.

Credit pricing

Base rate per million tokens: 400 credits for input, 2,000 for output The final cost is rounded up to whole credits.

Modes

claude-messages

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "claude-opus-5-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Claude Opus 4.8

claude-opus-4-8-chatAvailable via APIText input ready

Chat · Work through pasted text with Claude: ask for summaries, compare proposals, or revise a draft through follow-up chat.

Credit pricing

Base rate per million tokens: 400 credits for input, 2,000 for output The final cost is rounded up to whole credits.

Modes

claude-messages

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "claude-opus-4-8-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Claude Opus 4.7

claude-opus-4-7-chatAvailable via APIText input ready

Chat · Chat with Anthropic's Claude to summarize pasted text, compare options, and draft or revise written replies.

Credit pricing

Base rate per million tokens: 285 credits for input, 1,430 for output The final cost is rounded up to whole credits.

Modes

chat-completion

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "claude-opus-4-7-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Claude Opus 4.6

claude-opus-4-6-chatAvailable via APIText input ready

Chat · Send text messages and continue the conversation. Output is limited to 4096 tokens per reply.

Credit pricing

Base rate per million tokens: 285 credits for input, 1,430 for output The final cost is rounded up to whole credits.

Modes

claude-messages

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "claude-opus-4-6-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Claude Sonnet 4.6

claude-sonnet-4-6-chatAvailable via APIText input ready

Chat · Chat with Anthropic's Claude Sonnet 4.6 on Rivya. Paste text to request a summary, rewrite a draft or discuss code.

Credit pricing

Base rate per million tokens: 170 credits for input, 855 for output The final cost is rounded up to whole credits.

Modes

chat-completion

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "claude-sonnet-4-6-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Claude Opus 4.5

claude-opus-4-5-chatAvailable via APIText input ready

Chat · Send text messages and continue the conversation. Output is limited to 4096 tokens per reply.

Credit pricing

Base rate per million tokens: 285 credits for input, 1,430 for output The final cost is rounded up to whole credits.

Modes

claude-messages

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "claude-opus-4-5-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Claude Sonnet 4.5

claude-sonnet-4-5-chatAvailable via APIText input ready

Chat · Chat with Claude Sonnet 4.5 on Rivya. Paste a draft, an excerpt, or code, then ask for a rewrite, summary, or explanation.

Credit pricing

Base rate per million tokens: 170 credits for input, 855 for output The final cost is rounded up to whole credits.

Modes

chat-completion

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "claude-sonnet-4-5-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Claude Haiku 4.5

claude-haiku-4-5-chatAvailable via APIText input ready

Chat · A Claude text chat model on Rivya for summaries of pasted material, reply drafts, and working through everyday questions.

Credit pricing

Base rate per million tokens: 55 credits for input, 285 for output The final cost is rounded up to whole credits.

Modes

chat-completion

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "claude-haiku-4-5-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Grok 4.7

grok-4-7-chatThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Chat · Grok 4.7 targets coding, agent tasks and knowledge work, according to xAI’s guide. New requests are paused on Rivya.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Grok 4.6

grok-4-6-chatThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Chat · Grok 4.6 targets coding, agent tasks and knowledge work. New requests are paused on Rivya.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Grok 4.5

grok-4-5-chatThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Chat · Grok 4.5 targets coding, engineering and agent workflows. New requests are paused on Rivya.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Grok 4.3

grok-4-3-chatThis service is temporarily unavailable. Try again later.This service is temporarily unavailable. Try again later.

Chat · Grok 4.3 is described by xAI as a fast model for following instructions and calling tools. New requests are paused on Rivya.

Credit pricing

This service is temporarily unavailable. Try again later.

Modes

API input

This service is temporarily unavailable. Try again later.

Inputs

-

Prompt limit

-

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

No runnable Public API example is shown for this model.

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

You can browse this model for comparison, but it cannot currently be called through Public API v1. Check its API status before submitting a request.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Kimi K3.1 Pro

kimi-k3-1-pro-chatAvailable via APIText input ready

Chat · Chat using text, without attachments or web search. Credits are reserved before sending using a conservative output allowance; final billing uses confirmed consumption.

Credit pricing

Base rate per million tokens: 160 credits for input, 800 for output The final cost is rounded up to whole credits.

Modes

responses

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "kimi-k3-1-pro-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

Kimi K3

kimi-k3-chatAvailable via APIText input ready

Chat · Chat using text, without attachments or web search. Credits are reserved before sending using a conservative output allowance; final billing uses confirmed consumption.

Credit pricing

Base rate per million tokens: 480 credits for input, 2,400 for output The final cost is rounded up to whole credits.

Modes

responses

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "kimi-k3-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.

DeepSeek V4.1 Flash

deepseek-v4-1-flash-chatAvailable via APIText input ready

Chat · Chat using text, without attachments or web search. Credits are reserved before sending using a conservative output allowance; final billing uses confirmed consumption.

Credit pricing

Base rate per million tokens: 24 credits for input, 95 for output The final cost is rounded up to whole credits.

Modes

responses

API input

Text input ready

Inputs

text

Prompt limit

20000 characters

Reference media

-

Endpoint

POST /api/v1/chat/completions

Authentication

Bearer API key

Idempotency-Key

Use the same Idempotency-Key and body when retrying a request. A pending or changed request can return idempotency_conflict; check the original turn before starting a new request.

Request body

{
  "model": "deepseek-v4-1-flash-chat",
  "message": "Outline three steps for launching a product image campaign."
}

Reference rule

-

Result

Non-streaming Chat returns an assistant message, session_id, token usage, and credit usage with settlement status. Continue with session_id; do not poll the generation endpoint.

Common errors

validation_failed, insufficient_credits, idempotency_conflict, rate_limited, not_found, chat_model_not_supported, chat_session_conflict, chat_attachment_not_supported

Chat API supports non-streaming and SSE streaming turns in Public API v1. Image attachments are not supported for this chat model.

Parameters

No additional model controls are listed. Check reference-input requirements separately.