Integration

Minimal configuration

OpenAI-compatible clients usually use:

Base URL  https://dingjiai.com/v1
API Key   DINGJIAI user API Key
Model     The Chat Completions model-product public model id returned by GET /v1/models

Responses, Anthropic, and Gemini use their respective API/wire family entries. A model list for one API entry cannot prove that another API entry is available. Codex may use Search automatically during a Responses workflow; no extra client configuration is needed. See Base URL and authentication for the complete table.

Integration order

1. For Chat Completions, Anthropic, or Gemini, query that API family entry's public model list. Responses has no dedicated public model list; use a public model id explicitly enabled for the Responses API entry in the console or by an administrator, and do not infer availability from GET /v1/models.
2. Send one minimal request.
3. Save the X-Request-Id response header.
4. Confirm that usage, charges, and balance changes match expectations before connecting your business code.

For ready-made tools, see Codex, Claude Code, or One-click access.