Guide
Cursor with a custom OpenAI-compatible API
Cursor can override the OpenAI API. You need a base URL, a key and a published model slug — the same key can also serve your backend.
Updated
Settings
Open Settings → Models → OpenAI API Key. Enable the override, paste the LLMcode key and set the base URL to https://llmcode.io/v1.
Add the published profile as a custom model if Cursor asks for one. Use the slug from /models, not a display name.
Use a separate key
Cursor stores the key on the developer machine. Issue a dedicated key with a spend limit and expiry, so a leak does not affect production keys.
For autocomplete, choose the fastest published profile; for complex edits, switch to a larger one.
Troubleshooting
GET /v1/models with the same key shows what the key can see. 401 means a bad or revoked key; 403 means the model is not allowed for that key.
Cursor may cache the model list after a key change — restart it.
Published model prices
USD per 1M tokens. The numbers come from the public catalog on every page rebuild — this is the sale price, not an estimate.
| Model | Input | Output | Cache |
|---|---|---|---|
| Claude Opus 5.5 | $2.70 | $8.00 | $0.14 |
| DeepSeek Flash | $0.50 | $2.00 | $0.01 |
| Claude Sonnet 5.5 | $1.33 | $6.67 | $0.07 |
| DeepSeek V4 Pro | $2.20 | $6.60 | $0.07 |
Top-up from $10, balance in USD. 1 USD = 1,000,000 UT. No subscription and no auto-charge.
Questions
Does Cursor support Claude models through this?
It depends on the Cursor version and mode. Claude profiles are available through the compatible endpoint; check the model list the client receives.
How is Cursor usage billed?
Per token at the published model price. There is no separate editor fee.
Is a VPN required?
No VPN is required by us, but we do not promise to bypass network restrictions. Access depends on your network.
What does 429 mean?
The request rate limit was hit: 600 requests per minute by default. Retry after a pause or reduce autocomplete frequency.