Cursor is a native code editor built for AI programming. It deeply integrates full-project context awareness, smart code completion (Tab), sidebar chat, and a multi-file agent (Composer). Connect it with OpenRely API to effortlessly utilize top-tier models worldwide within Cursor.
Go to the Cursor official download page to get the latest installer for your operating system (supports macOS, Windows, and Linux), then complete the local installation and launch it.
[!TIP]
If you previously used VS Code, you can seamlessly import all your existing extensions, theme settings, and keybindings with one click upon launching Cursor for the first time.
Launch Cursor and press the general settings shortcut Cmd + , (Ctrl + , for Windows users), or click the gear settings icon ⚙️ in the top right corner of the interface;
In Settings, select Cursor Settings (or Models in the left navigation bar) to enter the model service management page.
Locate the OpenAI-compatible provider configuration section on the Models settings page:
Enter API Key: Fill in your OpenRely dedicated access token (sk-...) in the OpenAI API Key input field;
Override Base URL: Toggle on the Override OpenAI Base URL switch, and enter the OpenRely compatible endpoint:
https://api.openrely.ai/v1
(Cursor automatically appends /chat/completions to the end during requests; be sure to keep /v1 and do not add trailing slashes)
Verify Connection: Click the Verify button next to the input field. If a green checkmark or success message appears, network communication and authentication are working properly.
Cursor supports flexibly mounting any model supported by OpenRely:
Below the Model Names list, click + Add Custom Model;
Enter the model identifier you wish to use (such as claude, gpt, or deepseek; you can always check the OpenRely Model Pricing Center for the latest available model IDs);
After clicking save, ensure the toggle switch to the right of the model is enabled.
Return to the code editor window to invoke OpenRely-powered AI capabilities across different interfaces:
Sidebar Chat: Press Cmd + L (Windows: Ctrl + L), select your newly added custom model from the model switcher at the bottom right of the chat box, and ask it about code implementation or debugging advice;
Multi-file Agent (Composer): Press Cmd + I (Windows: Ctrl + I), enter your refactoring or development requirements, and let the AI autonomously locate files and generate a complete code diff.