# Quick Start: Cursor IDE Integration
This guide will help you integrate the MCP Router with Cursor IDE so it automatically selects the best model for each query.
## How It Works
```
Your Query → Router Analyzes → Best Model Selected → Cursor Uses That Model
```
## Step 1: Install Dependencies
```bash
cd mcp-router
pip install -r requirements.txt
```
## Step 2: No API Keys Needed!
**Important:** The router doesn't need API keys! It only recommends models to Cursor. Cursor uses its own configured API keys to make the actual API calls.
This means:
- ✅ No need to set API keys in the router
- ✅ Cursor already has your keys configured
- ✅ More secure - keys stay in Cursor
## Step 3: Run Setup Script
```bash
./scripts/setup_cursor.sh
```
This will:
- Find your Cursor config directory
- Configure the MCP server in `mcp.json`
## Step 4: Add to Agent Settings
After configuring the MCP server, you need to enable it in Cursor's agent settings:
### Option A: Via Settings UI
1. Open Cursor Settings (`Cmd+,` or `Ctrl+,`)
2. Go to `Features` → `Model Context Protocol`
3. Enable `mcp-router` server
4. Make sure "Use MCP Tools" is enabled in Agent settings
### Option B: Via Command Palette
1. Press `Cmd+Shift+P` (Mac) or `Ctrl+Shift+P` (Windows/Linux)
2. Type "Cursor: Open Agent Settings"
3. Enable MCP tools and select `mcp-router`
See `docs/AGENT_SETTINGS.md` for detailed instructions.
## Step 5: Verify Setup
This will:
- Find your Cursor config directory
- Create/update `mcp.json` with the router configuration
- Set up the MCP server path
## Step 4: Manual Configuration (Alternative)
If the script doesn't work, manually add to Cursor's MCP config:
**Location:**
- macOS: `~/Library/Application Support/Cursor/User/globalStorage/mcp.json`
- Linux: `~/.config/Cursor/User/globalStorage/mcp.json`
- Windows: `%APPDATA%\Cursor\User\globalStorage\mcp.json`
**Configuration:**
```json
{
"version": "1.0",
"mcpServers": {
"mcp-router": {
"command": "python3",
"args": [
"/absolute/path/to/mcp-router/src/mcp_server.py"
],
"env": {}
}
}
}
```
**Important:** Replace `/absolute/path/to/mcp-router/src/mcp_server.py` with the actual absolute path.
## Step 5: Restart Cursor
1. Close Cursor completely
2. Reopen Cursor
3. Check MCP status (should show "mcp-router" connected)
## Step 6: Test It!
1. Open Cursor chat
2. Type a query like: "Explain how neural networks work"
3. The router will automatically:
- Analyze your query
- Select the best model (e.g., Claude 3.5 Sonnet for reasoning)
- Cursor will use that model
## Verification
To verify it's working:
1. **Check MCP Status:**
- Open Cursor Settings → MCP
- You should see "mcp-router" listed and connected
2. **Test Routing:**
```bash
python3 main.py route "test query"
```
3. **Use MCP Tools in Cursor:**
- In Cursor chat, you can use MCP tools:
- `route_query` - Get routing recommendation
- `execute_routed_query` - Route and execute
- `list_models` - See available models
## Troubleshooting
### MCP Server Not Starting
1. **Check Python path:**
```bash
which python3
```
Make sure it's in your PATH
2. **Test MCP server directly:**
```bash
python3 src/mcp_server.py
```
Should start without errors
3. **Check file permissions:**
```bash
chmod +x src/mcp_server.py
```
### Models Not Available
1. **Check router:**
```bash
python3 main.py list
```
Should show all registered models
2. **Note:** Router doesn't need API keys - it just recommends models. Cursor handles API calls.
### Cursor Not Connecting
1. **Check config file exists:**
- Verify `mcp.json` is in the correct location
- Check JSON syntax is valid
2. **Check Cursor logs:**
- Look for MCP-related errors in Cursor's developer console
3. **Restart Cursor:**
- Fully quit and restart Cursor
## Example Queries
Try these to see different models selected:
- **Reasoning:** "Explain quantum computing principles"
→ Should route to Claude 3.5 Sonnet or GPT-4o
- **Code Generation:** "Write a Python function to sort a list"
→ Should route to GPT-4o-mini or Claude 3 Haiku (fast, cost-effective)
- **Complex Architecture:** "Design a microservices architecture"
→ Should route to GPT-4o or Claude 3.5 Sonnet (high quality)
## Next Steps
- Customize routing strategies
- Add more models
- Monitor routing statistics
- Fine-tune model selection criteria
## Support
For issues or questions:
1. Check `docs/cursor_integration.md` for detailed docs
2. Test router directly: `python3 main.py route "your query"`
3. Check MCP server logs in Cursor