Skip to main content

OpenAI

An llm · client with dialect: openai sends requests to the OpenAI Chat Completions API or the OpenAI Responses API. Requests that arrived in the Anthropic API are translated to the OpenAI format and back.

Configuration

openai_llm_client:
  type: llm
  kind: client
  options:
    dialect: openai
    server: https://api.openai.com
    authorization:
      api_key:
        credentials: "Bearer {credentials}"
  telemetry:
    metrics:
      - llm.*
  exit: sys:http_client

APIs

The openai dialect covers two OpenAI APIs:

APIRequest path
Chat Completions/v1/chat/completions
Responses/v1/responses

options.apis lists the APIs the provider supports, from chat/completions and responses. It defaults to both. The client calls the API of the request when it is listed. Otherwise it translates the request to the first listed API, so a Responses request can reach a server that has no /v1/responses:

openai_llm_client:
  type: llm
  kind: client
  options:
    dialect: openai
    apis:
      - chat/completions
    server: https://api.openai.com

A request from another dialect, such as Anthropic, is sent to the first listed API, which is Chat Completions when both are listed. See Dialect Translation for what is carried between the two APIs.

Server URL

Zilla appends /chat/completions or /responses, following the API in use, to the server URL. When the URL has no path, or only /, the path defaults to /v1, so https://api.openai.com sends Chat Completions requests to https://api.openai.com/v1/chat/completions.

The same applies to any server that implements the OpenAI Chat Completions or Responses API: set server to its base URL, and list the APIs it implements in apis.

Credentials

options.authorization adds the API key to each request as Authorization: Bearer <key>. With an inline guard, the key is the one the caller sent, forwarded unchanged. See LLM Credentials.

Routing to OpenAI

north_llm_proxy:
  type: llm
  kind: proxy
  routes:
    - when:
        - model:
            - gpt-*
      exit: openai_llm_client

Next Steps