↓
Rivya AI Docs

Rivya Data Processing, Storage and Deletion

Learn what Rivya saves, when reference uploads reach external services, and how provider terms, file availability and account deletion affect your data.

Last reviewed on September 13, 2026

Before submitting a prompt or uploading a file, consider both what Rivya needs to run your request and which external services may handle it. This guide explains the product behavior; the Privacy Policy and Terms of Service set out the applicable terms.

What does Rivya save?

Rivya may store submitted prompts, instructions, chat messages, reference-file URLs and upload metadata, model and tool selections, generation settings, task status, output URLs and result details. Chat records can include messages, attachments and usage information. Credit reservations, settlement and refunds also need records linked to the work.

These records support account access, generation, chat history, billing, support and investigation of errors or abuse. Technical information can include sessions, IP addresses, browser and device details, language preferences and security logs. Service and billing communications, and newsletters where applicable, use the information needed to deliver them.

Saved task or chat records do not mean every unfinished draft or setting is saved. An output URL in history is also different from a permanent copy of the file.

Which services may receive data?

The services involved depend on the feature you use and the integrations enabled for it. AI requests can pass through external AI infrastructure and the upstream model services selected for that request. A model's public page does not prove it processed your data or that it is available to run; check model availability before planning a task.

Other services may include Stripe for checkout, subscriptions and payment management; the configured email service for authentication, account, support or newsletter messages; and storage services for the uploads they handle. Google is involved when you choose Google sign-in. Legacy GitHub or Discord sign-in applies only if offered and selected. Hosting, security and monitoring services support operation; analytics, captcha, affiliate or support integrations apply only when configured and active.

Rivya does not store payment card details directly; Stripe handles them. The presence of an optional integration does not mean every user or request passes through it. For questions about a specific model's conditions, start with the provider and commercial-use table.

When is an uploaded file sent?

Reference files for supported models pass through Rivya's upload endpoints to the external AI infrastructure used for the request. Uploading a reference is a separate operation from submitting a generation or chat message. A file can therefore already have been sent even if you have not started generation or sent the message yet.

Review the file before uploading it. Removing a reference from the form is not confirmation that a remote copy has been erased. Avatar and other stored-file uploads use the storage service configured for that feature, so they should not be assumed to follow the same path as generation references.

Supported formats, sizes and reference counts vary by model. See references and uploads for the product limits; accepting a file does not establish permission to use its contents.

Is there a no-training or no-retention guarantee?

Do not infer a universal no-training or zero-retention guarantee from the fact that Rivya processes data to deliver a service. The public policy describes operational processing; it does not establish identical retention, review or secondary-use rules for every external provider.

If your project requires a specific restriction on training, review or retention, check the conditions for the actual processing path before sending content. If those conditions are unclear, resolve them first. Avoid sending sensitive personal information, confidential material or protected data unless you have assessed your requirements and can accept the external processing involved.

How do retention and account deletion work?

Rivya retains information for operational, billing, legal, security and support needs. Some successful media results have a 14-day availability window recorded by the application. That is not a universal deletion deadline for prompts, billing records, backups or provider-held data, nor a guarantee that an external link will remain accessible throughout that period. Follow output downloads to keep your own copies of files you need.

Deleting your account under Settings > Security generally removes account-linked records from Rivya's primary database through its deletion relationships. Limited records may remain where needed for billing, fraud prevention, abuse handling, legal obligations, backups, security logs or provider-side processing. Account deletion does not establish immediate erasure across every service.

Review account settings before deletion, including subscription handling: deleting an account is not the same as cancelling a subscription. For a privacy request you cannot handle in settings, use Contact.

What should you check before sending content?

Identify personal, confidential, client-owned or third-party material before uploading. Send only what the task needs; a summary may avoid exposing an entire source file, but check the summary itself for sensitive details. Confirm both the right to upload and the right to use the result, as explained in ownership and commercial use.

Use the safe upload guidelines for the practical checks. Reassess when you add real people, voices, unreleased product information or material that must stay within your organization. Permission to use a source, acceptance of the processing conditions and review of the finished result are separate decisions.