Cursor Integration Guide

Applicable roles: Developers, Admin Users, Member Users Last updated: 2026-08-06

Cursor's support for custom Base URLs varies across versions and account capabilities. Before configuring, open the model settings and confirm that the current interface provides both API Key and Base URL input fields.

Use the API endpoint shown on the console "API Key" page.

When the Current Version Provides a Base URL

1. Open Cursor Settings and navigate to Models.
2. Find the OpenAI API Key configuration.
3. Expand Override OpenAI Base URL.
4. Enter your platform API key in the API Key field.
5. Enter https://<your-api-endpoint> in the Base URL field.
6. Add the model call name from the console "Call Guide".
7. Save and select the model in the Chat panel.

If Verify does not pass but Chat returns responses normally and usage records appear in the console, rely on the actual request results. Some Cursor versions' Verify only recognizes built-in models, which does not necessarily mean custom models are unusable.

When the Current Version Does Not Provide a Base URL

Do not enter your platform API key into a key input field that can only connect to the official service. Instead, install a plugin that supports OpenAI Compatible in Cursor, such as Cline, Roo Code, or Continue, and follow the corresponding tutorial to configure it.

What Else to Check After Integration

Cursor's chat, custom models, and Tab completion may use different services. After configuration, verify each feature you actually need. Chat using the platform does not mean all built-in features will use the same Base URL.

If requests return 404, check whether the final request URL from Cursor contains a duplicated /v1. If a model is not found, copy the model name from the call guide again — do not use the model display name in place of the call name.