↓
Rivya AI Docs

Rivya API Models: Availability and Inputs

List Rivya models and check API availability, accepted inputs, parameter options, and Files API requirements before sending a request.

Last reviewed on November 11, 2026

Use GET /api/v1/models to find Rivya model IDs and check their current API availability and input requirements.

curl https://rivya.ai/api/v1/models

This endpoint is public and read-only; it does not require an API key. It includes callable models and entries listed only for discovery. The API reference and online debugger use this catalog, though the debugger only offers a subset of models. Finding a model in the list does not make it available for a request.

You can add the optional locale query parameter to request localized display names and parameter labels. Model IDs, parameter keys, option values, and status values remain unchanged. An unsupported locale falls back to English.

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.

Response Shape

This example shows selected fields for z-image. The live response also includes pricing and media-limit fields; read it for current values.

{
  "object": "list",
  "data": [
    {
      "id": "z-image",
      "display_name": "Z-Image",
      "category": "image",
      "runtime": "async_task",
      "base_credits": 1,
      "billing_type": "FIXED",
      "supported_modes": ["text-to-image"],
      "max_prompt_length": 1000,
      "reference_media": null,
      "params": [
        {
          "key": "aspect_ratio",
          "type": "select",
          "label": "Aspect ratio",
          "required": false,
          "default_value": "1:1",
          "options": [
            { "value": "1:1", "label": "1:1" },
            { "value": "4:3", "label": "4:3" },
            { "value": "3:4", "label": "3:4" },
            { "value": "16:9", "label": "16:9" },
            { "value": "9:16", "label": "9:16" }
          ]
        }
      ],
      "supports_direct_generation": true,
      "api_status": "available",
      "api_input_status": "available_text",
      "supported_api_inputs": ["text"],
      "requires_files_api": false,
      "requires_upload_for_modes": [],
      "url_input_supported": false,
      "file_kinds": [],
      "chat_capabilities": null,
      "notices": []
    }
  ]
}

For an available model with runtime: async_task, use its id as the model value in POST /api/v1/generations. For an available chat model, use POST /api/v1/chat/completions and check chat_capabilities.

Read the model fields

  • id is the model ID for API requests. Use it rather than display_name, which is a label for readers. category identifies the model category, and runtime distinguishes asynchronous generation from Chat.

  • supported_modes lists the workflows mapped in Rivya. It does not describe every feature offered by the provider. Follow the model reference for selecting a mode and supplying its inputs.

  • max_prompt_length describes a configured prompt limit. A null value does not mean that a prompt is unlimited or optional; check the model reference for its prompt requirements.

  • max_images, max_videos, max_audios, and reference_media describe configured media limits. A null limit does not mean uploads are unlimited. These fields do not describe every permitted combination or the minimum files required by a mode.

  • params lists model-specific parameter keys, types, labels, defaults, and options. Number parameters can include min, max, and step; text parameters can include max_length. Send option value strings rather than their display labels, and respect hide_when_modes when it is present.

  • A parameter marked required: false can still have conditions in a particular mode. Check the full model reference for required input combinations, file metadata, and other validation rules.

  • base_credits, billing_type, and pricing describe the current billing configuration. Final credits follow the model's billing and settlement rules; the example above is not a final charge for every request.

  • notices provides additional guidance about the model. Read it alongside the model reference and the relevant API documentation.

Availability States

Check api_status before choosing a model for a new request:

  • available: the model is eligible for Public API requests. Check supports_direct_generation, runtime, and the input requirements before submitting.

  • temporarily_unavailable: the model cannot currently be called through Public API v1. Choose an available model for a new request.

  • studio_only: this entry is not callable through Public API v1. Check its product page for current Studio availability.

  • coming_soon: the model is listed for discovery and cannot be submitted to generation or Chat API.

Entries that are not callable return null runtime, billing, pricing, prompt-limit, media-limit, and chat-capability fields, with empty supported_modes and params. Those empty values do not unlock a model or remove its input restrictions.

Input readiness

For an available model, api_input_status summarizes how to supply input. It does not replace the request's validation rules:

  • available_text: the model has a text-based input path. Depending on the model, the request uses prompt, params.dialogue, or a Chat message.

  • available_url_input: the model accepts URL input through its model parameters. Check the model's URL parameter and format requirements.

  • partial_requires_files: the model has text or URL input as well as file-based modes. Check requires_upload_for_modes and the model reference to see which inputs your mode requires.

  • requires_files_api: the model's input path depends on reference files. Upload the required media through /api/v1/files before submitting its request.

supported_api_inputs summarizes the input types as text, url, or file. file_kinds identifies accepted reference-media kinds; it does not mean that every kind can be combined in one request. Check the selected mode's requirements.

Use Files API before calling modes that require uploaded image, video, or audio references.

When a model requires signed media metadata, use the original model-bound Files API upload response. A URL alone, or metadata from a later GET /api/v1/files/{fileId}, may not include the tokens and verified fields needed for the request. The Files API documentation explains how to pass the upload fields.

url_input_supported: true describes a URL input path. It does not allow arbitrary URLs or waive file-format, ownership, duration, or signed-metadata checks required by the model.

For chat models, chat_capabilities describes non-streaming and SSE streaming responses, web search, image file_id attachments, thoughts, and supported reasoning efforts. Check the selected model's values before sending optional features; a feature supported by the provider may not be enabled in Rivya.

Model Count Policy

The public model list changes as Rivya adds, removes, or changes access to models.

Do not rely on a fixed model count in your integration. Read /api/v1/models or the Model API Reference for the current list.

The live OpenAPI schema lists current model IDs separately for new generation, Chat, and file-upload requests. A model listed by GET /api/v1/models is not necessarily included in every request enum.

Saved generation and Chat responses use a string model field so they can retain older IDs after availability changes. Do not reject a historical response just because its model is absent from a current request enum.

Coming Soon discovery entries

Coming Soon entries are for discovery, not early access. They return supports_direct_generation: false, null runtime, billing, and pricing fields, and empty supported_modes and params. Choose a callable model with api_status: available for a new request.

Questions about model availability

Why is a listed model absent from a request enum?

The public list includes unavailable and Coming Soon entries. Request enums include the models eligible for that endpoint, so generation, Chat, and file-upload requests have different model lists. Read the entry's availability and input fields before choosing an endpoint.

Does an empty notices array mean there are no restrictions?

No. notices is supplementary guidance, not a complete list of validation rules. Even an available model can require a particular file combination, format, or signed token. Follow the model reference and the endpoint documentation.