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

# API Overview

> Endpoints, formats, and conventions for the Openference /v1 API.

# API Overview

Openference exposes a stable OpenAI-compatible surface plus direct Anthropic format for Claude.

All requests require a valid `Authorization: Bearer YOUR_API_KEY` header.

## Base URL

```
https://api.openference.com/v1
```

## Core Endpoints

| Method & Path            | Format    | Description                             |
| ------------------------ | --------- | --------------------------------------- |
| `POST /chat/completions` | OpenAI    | Chat completions with streaming & tools |
| `POST /messages`         | Anthropic | Claude Messages API                     |
| `POST /responses`        | OpenAI    | Responses API (used by Codex etc.)      |
| `POST /embeddings`       | OpenAI    | Text embeddings                         |
| `GET /models`            | OpenAI    | List models available to the caller     |

## Format handling

* If the model route supports the client format natively, we pass through.
* Otherwise we automatically convert between OpenAI ↔ Anthropic ↔ Gemini.

See [Format conversion](/api-reference/format-conversion) for details.

## Model names

Use the exact model ID returned by `GET /v1/models` (or shown on the website).

Some routes may map an alias to a specific upstream model ID.

## Streaming

Add `"stream": true` (OpenAI) or the equivalent for Anthropic.

We forward SSE correctly.

## Errors

See [Errors](/api-reference/errors).

## Rate limits & quotas

See [Rate Limits](/api-reference/rate-limits).