> ## Documentation Index
> Fetch the complete documentation index at: https://docs.deepshi.ai/llms.txt
> Use this file to discover all available pages before exploring further.

> ## Agent Instructions
> The Deepshi API is an OpenAI-compatible gateway. Base URL: https://api.deepshi.ai/v1. API keys start with sk-bf- and go in the Authorization: Bearer header. Prefer the official OpenAI SDKs pointed at the Deepshi base URL. Model ids are clean with no provider prefix (e.g. deepshi-3.0, claude-opus-4.8, gpt-5.5). Chat and image requests are synchronous; video and music requests are asynchronous job APIs (create, then poll). Every synchronous response carries usage.cost.total_cost in USD. Do not reference the internal /api/* admin plane or virtual keys.

# Privacy

> How Deepshi handles your prompts and data.

Deepshi is built for people who want AI without surveillance. The principle is simple: we don't store the content of your prompts or responses, and we never use your data to improve models.

<Note>
  **Zero data retention.** Your prompt and response content is never stored, or used to train or
  fine-tune any model, Deepshi's or anyone else's.
</Note>

## Our own models

Deepshi's own models run on **private infrastructure that Deepshi operates**, and your requests are never handed to anyone else to serve. Your prompts and responses are never stored.

## Third-party models

When you call a model created by another provider, your request is processed only to generate the response. Deepshi maintains **zero-retention agreements** with the providers it routes to, so **your prompts and responses are not stored or used to train the provider's models**.

## How requests are handled

* All traffic travels over **HTTPS/TLS**.
* Requests are processed for **inference only**.
* Operational **metadata** (authentication, usage metering, billing, and abuse prevention) is processed to run the service. This is separate from your prompt and completion **content**.

## Logging & data retention

Deepshi does **not** store the content of your prompts and responses, anywhere or under any setting. There is no opt-in logging mode that changes this. Only usage metadata (tokens, cost, latency, model, and status) is recorded, so you can review your spend and history in the dashboard.