# `ReqLLM.Providers.OpenAI.API`
[🔗](https://github.com/agentjido/req_llm/blob/main/lib/req_llm/providers/openai/api.ex#L1)

Behaviour for OpenAI API endpoint drivers.

Defines the contract for modules that implement OpenAI API-specific request/response handling.
The OpenAI provider uses this behaviour to support multiple API endpoints with different
request/response formats.

## Implementations

- `ReqLLM.Providers.OpenAI.ChatAPI` - Chat Completions API (`/v1/chat/completions`)
- `ReqLLM.Providers.OpenAI.ResponsesAPI` - Responses API (`/v1/responses`)
- `ReqLLM.Providers.OpenAI.ImagesAPI` - Images API (`/v1/images/generations`, `/v1/images/edits`)

## Callbacks

- `path/0` - Returns the default API endpoint path
- `path/1` - Optionally returns a non-default API endpoint path (e.g. edits endpoint for images)
- `encode_body/1` - Transforms request into provider-specific JSON format
- `decode_response/1` - Parses API responses into ReqLLM structures
- `decode_stream_event/2` - Decodes server-sent events for streaming
- `attach_stream/4` - Builds Finch streaming request with proper headers/body

# `attach_stream`

```elixir
@callback attach_stream(
  LLMDB.Model.t(),
  ReqLLM.Context.t(),
  keyword(),
  atom()
) :: {:ok, Finch.Request.t()} | {:error, Exception.t()}
```

# `attach_websocket_stream`
*optional* 

```elixir
@callback attach_websocket_stream(LLMDB.Model.t(), ReqLLM.Context.t(), keyword()) ::
  {:ok, map()} | {:error, Exception.t()}
```

# `decode_response`

```elixir
@callback decode_response({Req.Request.t(), Req.Response.t()}) ::
  {Req.Request.t(), Req.Response.t() | Exception.t()}
```

# `decode_stream_event`

```elixir
@callback decode_stream_event(map(), LLMDB.Model.t()) :: [ReqLLM.StreamChunk.t()]
```

# `encode_body`

```elixir
@callback encode_body(Req.Request.t()) :: Req.Request.t()
```

# `path`

```elixir
@callback path() :: String.t()
```

# `path`
*optional* 

```elixir
@callback path(:edit) :: String.t()
```

---

*Consult [api-reference.md](api-reference.md) for complete listing*