1. Check Cursor's current Models settings
Use only settings that your Cursor build documents or exposes. Avoid browser extensions, patched clients or hidden unsupported overrides for account-bound editor traffic.
Cursor and custom providers
Use TKEN where your coding workflow supports a base URL
Cursor has documented BYOK limits, so the safest approach is to separate Cursor-native model settings from OpenAI-compatible workflows that explicitly support a custom base URL.
Gateway smoke test
server-side
curl https://www.tken.shop/v1/chat/completions \
-H "Authorization: Bearer $TKEN_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "MiniMax-M2.7",
"messages": [
{"role": "user", "content": "Summarize this code change."}
]
}'Respect Cursor documented BYOK limits
Use TKEN in tools with custom base URL support
Keep gateway tests outside editor settings
Recommended routes
If your Cursor version or account exposes only official provider key fields, do not force unsupported behavior. Put TKEN in the surrounding developer stack where a custom OpenAI-compatible base URL is first-class.
Base URL
https://www.tken.shop/v1
Best fit
SDK, LiteLLM, Continue, Open WebUI
Cursor caution
Custom keys are limited by Cursor
Secret handling
Use environment variables
Safe setup decision tree
2. Run TKEN in an explicit custom-provider tool
For coding-agent experiments, use Continue or LiteLLM when you need a custom OpenAI-compatible base URL.
3. Keep a repeatable SDK smoke test
Use Node.js, Python or cURL to verify the TKEN key and model name before changing any editor-level configuration.
TKEN is an independent third-party API gateway. It is not affiliated
with Cursor or with OpenAI, Anthropic, Google, DeepSeek, Qwen, Kimi,
MiniMax, GLM or other model providers.
Use the gateway where base URLs are supported
Start with SDK and proxy workflows, then keep Cursor-specific settings inside the limits Cursor documents for your account.