Cursor supports OpenAI-compatible endpoints with a custom base URL.
-
Open Cursor Settings -> Models.
-
Under OpenAI API Key, paste your Talos key (
talos_...). -
Enable Override OpenAI Base URL and set it to:
https://api.usetalos.xyz/v1(In development:
http://localhost:8080/v1.) -
In the model list, add a custom model named exactly:
talos-autoYou can also add open models like
talos-llama-3.1-8b. -
Click Verify. Select
talos-autoin the chat/model picker.
See talos.env for the three values above in one place to
copy from.
Run verify.sh to confirm your key and the gateway work
before wiring up Cursor:
export TALOS_API_KEY=talos_YOUR_KEY
bash verify.shFor a runnable client instead of curl, see openai_sdk.py
(same script as examples/python/).
Notes:
- Cursor sends OpenAI
chat/completionsrequests; the gateway streams responses back in the same format. - If Verify fails, confirm the key is active on your dashboard and the base URL
ends with
/v1.