Setup time: ~2 minutes
Prerequisites
- Cursor editor installed
- Infercom API key
- OpenAI API key (required by Cursor to activate)
Configuration
Step 1: Open Model Settings
- Open Cursor
- Go to Settings (gear icon in left panel)
- Click on Models
Step 2: Configure API Endpoint
- Scroll to the API Keys section
- Enable Override OpenAI Base URL
- Enter:
https://api.infercom.ai/v1 - Enter your Infercom API key in the OpenAI API Key field
- Also enter a valid OpenAI API Key (required by Cursor)
Step 3: Add Custom Model
- Click + Add Custom Model
- Enter the model name:
MiniMax-M2.5
Step 4: Select Your Model
- In the models list, disable all default models
- Enable your Infercom model
- Select
MiniMax-M2.5in the right panel when coding
Model
UseMiniMax-M2.5 - optimized for agentic coding with 160K context (163,840 tokens) and 75.8% SWE-bench.
Usage
Once configured, Cursor will route all AI requests through Infercom:- Inline completion: Type and see suggestions
- Chat: Use
Cmd+L/Ctrl+Lto open chat - Edit: Select code and use
Cmd+K/Ctrl+K
Troubleshooting
AI Features Not Working
Ensure you have both:- A valid OpenAI API key entered (activates Cursor’s AI features)
- Your Infercom API key in the Override field
Model Not Responding
Verify the endpoint is correct:Wrong Model Selected
Check that:- Default models are disabled
- Your custom model is enabled
- The correct model is selected in the chat panel
Limitations
- Requires valid OpenAI API key to activate editor
- Some Cursor-specific features may not work (Cursor Tab, etc.)
- Model names show as custom rather than branded
Next Steps
- Cline - VS Code extension alternative
- Continue - Open-source VS Code extension
- Choosing a Tool - Compare all options