Integrations
Cursor
Use ai.ml models in Cursor's chat through its OpenAI base URL override.
Cursor can send its chat requests to an OpenAI-compatible API in place of OpenAI's. In Cursor Settings → Models:
- Paste your ai.ml key into OpenAI API Key and turn on Use OpenAI API key. Pasting the key alone does nothing.
- Turn on Override OpenAI Base URL and enter
https://api.ai.ml/v1. - Under Custom Models, add each ai.ml model id you want, such as
anthropic/claude-sonnet-4.6.
What to know:
- Only chat uses this key. Tab completion stays on Cursor's own models.
- Cursor sends these requests from its own servers, so the base URL must be reachable from the internet.
- With the override on, every OpenAI-family model in Cursor goes to ai.ml, including Cursor's built-in GPT models. Turn the override off to use those again.
- Cursor's agent can offer GPT models a freeform
apply_patchtool. ai.ml does not support freeform tools yet and answersunsupported_feature.custom_tools. Use the agent with a non-GPT model, or use chat. - Cursor's zero-data-retention promise covers its own models. For requests on your key, your ai.ml data policy applies.
Checked against Cursor's documentation and staff answers on its forum on 2026-10-02; not run here.