After connecting Cursor to Nbility, you can use GPT, Claude, and other models directly in the editor for chat, code edits, error explanations, and assisted development.
If you do not have an account or token yet, read Getting Started first.
1. Prepare
Before you begin, confirm that:
- You have registered an Nbility account
- You have signed in to the Nbility console
- You have created an active token
- The token has enough balance
If these are not ready, go to the registration page and the console first.
2. Open Cursor Settings
Launch Cursor and open the settings page.
If this is your first setup, search for Models so you can quickly find the
model and API options.
3. Open the Models Configuration
Find the Models section in Cursor settings.
This page usually includes model toggles, OpenAI API Key, Base URL, and related settings. Nbility integration is mainly configured here.
4. Fill in API Information
In the OpenAI-related settings, enter:
OpenAI API Key: the token you created in the Nbility consoleOverride OpenAI Base URL:https://api.nbility.dev/v1
If Cursor asks you to enable custom models or an OpenAI-compatible endpoint, follow the prompt.
5. Choose Models
After adding the key and Base URL, enable the models you need in Cursor.
Common choices:
- GPT series: prepare a
codexor related GPT group token - Claude series: prepare a
claudeorclaude-mixgroup token - Gemini: prepare the corresponding Gemini group token
If the current token belongs to only one group, it can only access models supported by that group. This is normal and does not mean the setup is wrong.
6. Verify the Integration
After configuration, run a simple verification immediately:
- Create a new chat or open any project
- Select an enabled model in Cursor
- Send a simple prompt, such as "Hello, please introduce yourself"
- Check whether a response returns normally
If you receive a normal response, Cursor is connected to Nbility.
7. Common Issues
1. Cursor Says the Key Is Invalid
Check these first:
- Whether the token was copied completely
- Whether extra spaces were pasted
- Whether the token has been deleted or disabled
Create a new token from the console tokens page and test again.
2. The Key Saves, but Requests Fail
Check these first:
- Whether the Base URL is exactly
https://api.nbility.dev/v1 - Whether your network or proxy affects Cursor requests
- Whether the token has enough balance
If a proxy app is running on your computer, quit it first, restart Cursor, and try again.
3. The Model I Want Is Missing
This usually means the current token group does not support that model.
Create a token for the corresponding group, then switch back to Cursor and use the new token.
8. Tips
- Create a separate token for Cursor to track usage
- In the beginning, use one group to verify the flow before configuring many models
- If you often switch models, split tokens by tool or use case
After this, you can also read: