OpenAI API Quickstart

CODEX RELAY / OPENAI GATEWAY

Use one API key for OpenAI Responses, Chat Completions, Embeddings, and Images. Follow this guide to list your models and send a first request in a few minutes.

OpenAI CompatibleResponses FirstDEV-0807

API Base URLhttps://codexapi.asia

POSTResponses API/v1/responsesPreferred by Codex and modern OpenAI clientsPOSTChat Completions/v1/chat/completionsBroad support across chat, IDE, and desktop clientsPOSTEmbeddings/v1/embeddingsVector search, retrieval, and knowledge workflowsPOSTImages API/v1/images/generationsImage generation, editing, and Base64 results

1. Create an API key

Sign in to the console

Open the Codex Relay console. Create an account first if registration is currently available.

Open the API Keys page

Create a key, select a group that publishes the OpenAI models you need, and give it a recognizable name.

Store it immediately

Treat the key like a password. Do not place it in chat history, screenshots, browser code, or public repositories.

2. List available models

Replace YOUR_API_KEY with the key you created:

curl https://codexapi.asia/v1/models \
  -H "Authorization: Bearer YOUR_API_KEY"

The key's assigned group determines the model list. A client model name must exactly match an id returned by this request.

3. Send your first request

This example uses the OpenAI-compatible Chat Completions endpoint. Replace MODEL_ID with a model from the previous response:

curl https://codexapi.asia/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "MODEL_ID",
    "messages": [
      {"role": "user", "content": "Say hello in one sentence."}
    ],
    "stream": false
  }'

Which Base URL should I use?

• For an API Host/Base URL field, start with https://codexapi.asia.
• For the OpenAI SDK base_url, use https://codexapi.asia/v1.
• Make sure the final request path does not contain /v1/v1/....

Next steps

Connect CodexCodex++ and config.toml configurationUse the OpenAI SDKPython, Node.js, and curl examplesConfigure desktop clientsStandard OpenAI-compatible provider settingsTroubleshoot401, 404, 429, 503, and missing-image issues