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

# API overview

> Integrate Mibyan through an OpenAI-compatible text API

Use the canonical base URL:

```text theme={null}
https://api.mibyanai.com/v1
```

The API is project-scoped. A request is authorized by the project key, evaluated against that project's model and endpoint permissions, and recorded in the project's usage and logs. Keep keys on a server and treat prompts and outputs as application data with the retention and privacy controls appropriate to your use case.

## Supported endpoints

| Method | Path                | Purpose                                      |
| ------ | ------------------- | -------------------------------------------- |
| `GET`  | `/models`           | List models available to the project         |
| `POST` | `/chat/completions` | Chat Completions with optional SSE streaming |
| `POST` | `/responses`        | Text and message Responses                   |

Use the public model ID `mibyan-4.1` for new integrations. The legacy `mibyan` alias may remain accepted for existing clients, but new integrations should use the current ID returned by `GET /v1/models`. Embeddings, images, audio, and file endpoints are not currently exposed.

<Card title="Platform quickstart" icon="rocket" href="/platform/quickstart">
  Create a project key and send a verified request.
</Card>

## Request lifecycle

1. Your server sends an HTTPS request with a bearer project key.
2. Mibyan authenticates the key and checks project status, permissions, model access, budget, and rate limits.
3. The request is passed to the active Mibyan model configuration.
4. The API returns a JSON response or SSE events and includes a request correlation ID.
5. Usage, latency, status, and error information can be inspected in Platform logs.

Use the documented endpoint schema and response behavior for your integration.