> ## Documentation Index
> Fetch the complete documentation index at: https://copilot-api.nick3.top/llms.txt
> Use this file to discover all available pages before exploring further.

# OpenAI-compatible API endpoints reference

> Reference for Copilot API's OpenAI-compatible endpoints: chat completions, responses, models, and embeddings — drop-in compatible with the OpenAI SDK.

These endpoints mirror the OpenAI API so any tool, SDK, or library built for OpenAI works with Copilot API without modification. Point your existing OpenAI client at `http://localhost:4141` and your requests go to GitHub Copilot instead.

## Base URL

```
http://localhost:4141
```

Use a different base URL if you started the server with a custom `--port`.

## Authentication

When you have API keys configured in `auth.apiKeys`, pass one of the following headers with every request:

| Header          | Format                        |
| --------------- | ----------------------------- |
| `x-api-key`     | `x-api-key: <key>`            |
| `Authorization` | `Authorization: Bearer <key>` |

If `auth.apiKeys` is empty, the server accepts requests without authentication.

***

## POST /v1/chat/completions

Creates a model response for a chat conversation in OpenAI Chat Completions format. Compatible with any tool that uses the OpenAI SDK or calls the OpenAI API directly.

<CodeGroup>
  ```bash curl theme={null}
  curl http://localhost:4141/v1/chat/completions \
    -H "Content-Type: application/json" \
    -d '{
      "model": "gpt-5.4",
      "messages": [{"role": "user", "content": "Hello!"}],
      "stream": true
    }'
  ```

  ```python openai-sdk theme={null}
  from openai import OpenAI

  client = OpenAI(
      base_url="http://localhost:4141/v1",
      api_key="your_key_or_dummy"
  )

  response = client.chat.completions.create(
      model="gpt-5.4",
      messages=[{"role": "user", "content": "Hello!"}],
      stream=True
  )
  ```
</CodeGroup>

***

## POST /v1/responses

OpenAI's most advanced interface for generating model responses (Responses API). Used by Claude Code and OpenCode for complex agent workflows, including multi-turn tool use and session continuations.

When `useResponsesApiWebSearch` is enabled in your `config.json` (the default), requests that include a `web_search` tool are forwarded upstream with the tool intact. Set `useResponsesApiWebSearch: false` to strip web search tools before sending.

```bash theme={null}
curl http://localhost:4141/v1/responses \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.4",
    "input": [{"role": "user", "content": "Explain async/await in JavaScript."}]
  }'
```

***

## GET /v1/models

Lists all models available from your Copilot account in OpenAI format.

```bash theme={null}
curl http://localhost:4141/v1/models
```

Example response:

```json theme={null}
{
  "object": "list",
  "data": [
    { "id": "gpt-5.4", "object": "model" },
    { "id": "gpt-5-mini", "object": "model" },
    { "id": "claude-opus-4-6", "object": "model" }
  ]
}
```

<Note>
  Actual model availability depends on your GitHub Copilot subscription tier. The list returned reflects what your account can access at the time of the request.
</Note>

***

## POST /v1/embeddings

Creates an embedding vector representing the input text.

```bash theme={null}
curl http://localhost:4141/v1/embeddings \
  -H "Content-Type: application/json" \
  -d '{
    "model": "text-embedding-3-small",
    "input": "The quick brown fox jumps over the lazy dog"
  }'
```

Example response:

```json theme={null}
{
  "object": "list",
  "data": [
    {
      "object": "embedding",
      "index": 0,
      "embedding": [0.0023, -0.0098, ...]
    }
  ],
  "model": "text-embedding-3-small",
  "usage": {
    "prompt_tokens": 9,
    "total_tokens": 9
  }
}
```