Cursor2API

English | 简体中文

A Go service that converts Cursor Web into an OpenAI chat/completions compatible API for local deployment.

✨ Features

• ✅ Compatible with OpenAI chat/completions
• ✅ Supports streaming and non-streaming responses
• ✅ High-performance Go implementation
• ✅ Automatic Cursor Web authentication
• ✅ Clean web interface
• ✅ Supports tools, tool_choice, and tool_calls
• ✅ Automatically derives -thinking public models
• ❌ Does not yet support Anthropic /v1/messages or MCP

🖼️ Screenshots

Drop images into docs/images/ and the README will render them.

🤖 Supported Models

• Google Gemini: gemini-3-flash
• Derived thinking model: gemini-3-flash-thinking

🚀 Quick Start

Requirements

• Go 1.24+
• Node.js 18+ (for JavaScript execution)

Local Running Methods

Method 1: Direct Run (Recommended for development)

Linux/macOS:

git clone https://github.com/libaxuan/cursor2api-go.git
cd cursor2api-go
chmod +x start.sh
./start.sh

Windows:

# Double-click or run in cmd
start-go.bat

# Or in Git Bash / Windows Terminal
./start-go-utf8.bat

Method 2: Manual Compile and Run

# Clone the project
git clone https://github.com/libaxuan/cursor2api-go.git
cd cursor2api-go

# Download dependencies
go mod tidy

# Build
go build -o cursor2api-go

# Run
./cursor2api-go

Method 3: Using go run

git clone https://github.com/libaxuan/cursor2api-go.git
cd cursor2api-go
go run main.go

The service will start at http://localhost:8002

🚀 Server Deployment Methods

Docker Deployment

1. Build Image:

# Build image
docker build -t cursor2api-go .

1. Run Container:

# Run container (recommended)
docker run -d \
  --name cursor2api-go \
  --restart unless-stopped \
  -p 8002:8002 \
  -e API_KEY=your-secret-key \
  -e DEBUG=false \
  cursor2api-go

# Or run with default configuration
docker run -d --name cursor2api-go --restart unless-stopped -p 8002:8002 cursor2api-go

Docker Compose Deployment (Recommended for production)

1. Using docker-compose.yml:

# Start service
docker-compose up -d

# Stop service
docker-compose down

# View logs
docker-compose logs -f

1. Custom Configuration: Modify the environment variables in the docker-compose.yml file to meet your needs:

• Change API_KEY to a secure key
• Adjust MODELS, TIMEOUT, and other configurations as needed
• Change the exposed port

System Service Deployment (Linux)

1. Compile and Move Binary:

go build -o cursor2api-go
sudo mv cursor2api-go /usr/local/bin/
sudo chmod +x /usr/local/bin/cursor2api-go

1. Create System Service File /etc/systemd/system/cursor2api-go.service:

[Unit]
Description=Cursor2API Service
After=network.target

[Service]
Type=simple
User=your-user
WorkingDirectory=/home/your-user/cursor2api-go
ExecStart=/usr/local/bin/cursor2api-go
Restart=always
Environment=API_KEY=your-secret-key
Environment=PORT=8002

[Install]
WantedBy=multi-user.target

1. Start Service:

# Reload systemd configuration
sudo systemctl daemon-reload

# Enable auto-start on boot
sudo systemctl enable cursor2api-go

# Start service
sudo systemctl start cursor2api-go

# Check status
sudo systemctl status cursor2api-go

📡 API Usage

List Models

curl -H "Authorization: Bearer 0000" http://localhost:8002/v1/models

Non-Streaming Chat

curl -X POST http://localhost:8002/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer 0000" \
  -d '{
    "model": "gemini-3-flash",
    "messages": [{"role": "user", "content": "Hello!"}],
    "stream": false
  }'

Streaming Chat

curl -X POST http://localhost:8002/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer 0000" \
  -d '{
    "model": "gemini-3-flash",
    "messages": [{"role": "user", "content": "Hello!"}],
    "stream": true
  }'

Tool Request

curl -X POST http://localhost:8002/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer 0000" \
  -d '{
    "model": "gemini-3-flash",
    "messages": [{"role": "user", "content": "Check the weather in Beijing"}],
    "tools": [
      {
        "type": "function",
        "function": {
          "name": "get_weather",
          "description": "Get current weather",
          "parameters": {
            "type": "object",
            "properties": {
              "city": {"type": "string"}
            },
            "required": ["city"]
          }
        }
      }
    ]
  }'

-thinking Model

curl -X POST http://localhost:8002/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer 0000" \
  -d '{
    "model": "gemini-3-flash-thinking",
    "messages": [{"role": "user", "content": "Think first, then decide whether a tool is needed"}],
    "tools": [
      {
        "type": "function",
        "function": {
          "name": "lookup",
          "parameters": {
            "type": "object",
            "properties": {
              "q": {"type": "string"}
            },
            "required": ["q"]
          }
        }
      }
    ],
    "stream": true
  }'

Use in Third-Party Apps

In any app that supports custom OpenAI API (e.g., ChatGPT Next Web, Lobe Chat):

1. API URL: http://localhost:8002
2. API Key: 0000 (or custom)
3. Model: Choose a supported base model or its automatically derived -thinking variant

⚙️ Configuration

Environment Variables

Variable	Default	Description
PORT	8002	Server port
DEBUG	false	Debug mode (shows detailed logs and route info when enabled)
API_KEY	0000	API authentication key
MODELS	gemini-3-flash	Base model list (comma-separated); the service automatically exposes matching -thinking public models
TIMEOUT	60	Request timeout (seconds)
KILO_TOOL_STRICT	false	Kilo Code compatibility: if tools are provided and tool_choice=auto, treat it as “tool use required”

Debug Mode

By default, the service runs in clean mode. To enable detailed logging:

Option 1: Modify .env file

DEBUG=true

Option 2: Use environment variable

DEBUG=true ./cursor2api-go

Debug mode displays:

• Detailed GIN route information
• Verbose request logs
• x-is-human token details
• Browser fingerprint configuration

Troubleshooting

Having issues? Check the Troubleshooting Guide for solutions to common problems, including:

• 403 Access Denied errors
• Token fetch failures
• Connection timeouts
• Cloudflare blocking

🧩 Kilo Code / Agent Orchestrator Compatibility

Some orchestrators enforce “must use tools” and may throw errors like MODEL_NO_TOOLS_USED when a response contains no tool call.

• Recommended: set KILO_TOOL_STRICT=true in .env
• Non-stream safety net: if tools are provided and tool use is required (tool_choice=required/function, or KILO_TOOL_STRICT), but the first attempt produces no tool_calls, the server automatically retries once (non-stream only)

Windows Startup Scripts

Two Windows startup scripts are provided:

• start-go.bat (Recommended): GBK encoding, perfect compatibility with Windows cmd.exe
• start-go-utf8.bat: UTF-8 encoding, for Git Bash, PowerShell, Windows Terminal

Both scripts have identical functionality, only display styles differ. Use start-go.bat if you encounter encoding issues.

🧪 Development

Running Tests

# Run existing tests
go test ./...

Building

# Build executable
go build -o cursor2api-go

# Cross-compile (e.g., for Linux)
GOOS=linux GOARCH=amd64 go build -o cursor2api-go-linux

📁 Project Structure

cursor2api-go/
├── main.go              # Main entry point (Go version)
├── config/              # Configuration management (Go version)
├── handlers/            # HTTP handlers (Go version)
├── services/            # Business service layer (Go version)
├── models/              # Data models (Go version)
├── utils/               # Utility functions (Go version)
├── middleware/          # Middleware (Go version)
├── jscode/              # JavaScript code (Go version)
├── static/              # Static files (Go version)
├── start.sh             # Linux/macOS startup script
├── start-go.bat         # Windows startup script (GBK)
├── start-go-utf8.bat    # Windows startup script (UTF-8)

└── README.md            # Project documentation

🤝 Contributing

Contributions are welcome! Please follow these steps:

1. Fork the repository
2. Create a feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'feat: Add some AmazingFeature')
4. Push to the branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

Code Standards

• Follow Go Code Review Comments
• Format code with gofmt
• Check code with go vet
• Follow Conventional Commits for commit messages

📄 License

This project is licensed under PolyForm Noncommercial 1.0.0. Commercial use is not permitted. See the LICENSE file for details.

⚠️ Disclaimer

Please comply with the terms of service of related services when using this project.

---

⭐ If this project helps you, please give us a Star!