Cursor

Connect Cursor to APIClaw with OpenAI API Key override fields, then add a model ID from the model hub.

Prerequisites

1. Create an API key in the API Keys dashboard.
2. Open the model hub and copy the full model ID your account can use.

Configure Cursor

Copy this Base URL first (you will paste it in step 3):

https://apiclaw.biz/v1

1. Open Cursor Settings and go to Models.
2. Find OpenAI in the provider list, paste your APIClaw API key into the field, and click Save.
3. Turn on Override OpenAI Base URL and paste the Base URL you copied above.
4. Click + Add Custom Model and paste an apiclaw/ model ID from the model hub, for example apiclaw/claude-fable-5 or apiclaw/gpt-5.6-sol.
5. Disable the built-in models in the list so only your APIClaw model is enabled.
6. Click Verify. Cursor only checks that the endpoint answers; it does not check that every mode works.
7. Turn Auto off, select your model in chat, and send a test message.

Use the apiclaw/ spelling of the model name. Cursor ships models of its own called gpt-5.6-sol and claude-fable-5-1, and a custom model whose name collides with one of those is refused; the prefixed id avoids the collision and bills and behaves identically.

While the base URL override is on, Cursor sends its built-in models to that URL too, which is why step 5 matters.

For agent mode, use the Cursor endpoint

Cursor's agent sends a Responses-shaped body to the chat completions path, which most gateways reject. APIClaw serves a Cursor-specific endpoint that accepts both shapes, so if agent mode misbehaves, set the Base URL to:

https://apiclaw.biz/cursor

Everything else stays the same, and requests are billed and logged exactly as on the normal base URL.

What the custom key covers

Chat routes to your custom base URL. Tab completion always stays on Cursor's own models whatever you configure. Agent and Edit are gated by Cursor and may answer "cannot be billed to an API key", and Composer models refuse a custom key outright. Cursor's zero-data-retention policy also does not apply to requests made with your own key, and Teams and Enterprise plans still charge Cursor's own per-token rate on top of APIClaw.

Verify

Send a short message in Cursor chat with your APIClaw model selected. If the model responds, the connection is working. You can also confirm the request in logs.

Troubleshooting

This model does not support custom API keys

Cursor blocks its own models from custom keys. Add the model with its apiclaw/ prefix, turn Auto off, disable the built-in models, and select the APIClaw model explicitly. Composer models cannot be used with a custom key at all.

Model not available in chat

Disable Auto mode and manually select the custom model you added from the model hub.

Unauthorized User API key or provider connection errors

Check:

• API key is correct and active in the API Keys dashboard
• Base URL is exactly https://apiclaw.biz/v1 (include /v1, do not append /chat/completions)
• Model ID was copied in full from the model hub
• Your plan still has available daily requests