An open-source implementation of the Cursor Agent core workflow, making it easy for everyone to learn and use. This project demonstrates the fundamental processes behind Cursor Agent functionality.
🎯 Special Note: This entire project, including this README.md, was automatically generated by Cursor. The programmer only used the chat panel without writing a single line of code in the editor. This showcases the power of AI-driven development, similar to how CLI tools like Claude Code are setting new standards for developer productivity.
🔧 Technical Details: The prompts and tools used in this codebase are replicated from Cursor's official native tools, providing an authentic reproduction of the original Cursor Agent workflow.
- 🚀 Open Source: Complete implementation of Cursor Agent core workflow
- 🤖 AI Integration: Default integration with DeepSeek AI model
- 🔧 Easy Configuration: Simple environment variable setup
- 🌐 API Compatibility: Support for OpenAI API compatible models
- 📦 Portable: Single binary executable
- 💡 Educational: Perfect for learning AI agent workflows
See openCursor in action! Here's how you can create a complete Snake game with a single command:
openCursor "Write a Snake game in HTML with JavaScript. Include proper game controls, scoring system, and game over functionality. Make it responsive and visually appealing."This command will generate a complete, playable Snake game including:
- ✅ HTML structure with canvas
- ✅ JavaScript game logic
- ✅ Keyboard controls (WASD or Arrow keys)
- ✅ Score tracking
- ✅ Game over detection
- ✅ Responsive design
- ✅ Modern styling
The AI agent will create all necessary files and provide you with a fully functional game that you can run immediately in your browser!
Download the openCursor binary and place it in your system's bin directory:
# Download the binary (replace with actual download URL)
wget https://github.com/your-username/openCursor/releases/latest/download/openCursor
# Make it executable
chmod +x openCursor
# Move to system bin directory
sudo cp openCursor /usr/local/bin/
# Now you can use it from anywhere
openCursor --helpMethod 1: Environment Variables
If you're using zsh, add to ~/.zshrc:
export OPENAI_API_KEY="your-deepseek-api-key"If you're using bash, add to ~/.bashrc:
export OPENAI_API_KEY="your-deepseek-api-key"Method 2: Custom Model Configuration
You can also configure other OpenAI API compatible models:
export OPENAI_API_KEY="your-api-key"
export MODEL="deepseek-chat"
export BASE_URL="https://api.deepseek.com/v1"# Basic usage
openCursor "Hello, please introduce yourself"
# Programming assistance
openCursor "Write a Python function to calculate Fibonacci sequence"
# Code review
openCursor "Review this code for potential improvements"- Visit DeepSeek Open Platform
- Sign up and log in to your account
- Navigate to API Keys management page
- Create a new API key
- Copy the key and configure it in your environment
📖 For detailed API documentation and usage examples, visit: DeepSeek API Documentation
- Go 1.19+: Make sure you have Go 1.19 or higher installed
- Git: For cloning the repository
- Make: For using the Makefile (optional)
Check your Go version:
go version# Clone the repository
git clone https://github.com/your-username/openCursor.git
cd openCursor
# Download dependencies
go mod download
# Build the project
go build -o openCursor
# Or use Make
make buildStandard Build:
go build -o openCursorOptimized Build (smaller binary):
go build -ldflags="-s -w" -o openCursorDebug Build:
go build -gcflags="all=-N -l" -o openCursor-debugBuild for different operating systems and architectures:
# Windows 64-bit
GOOS=windows GOARCH=amd64 go build -o openCursor.exe
# macOS 64-bit (Intel)
GOOS=darwin GOARCH=amd64 go build -o openCursor-macos-intel
# macOS 64-bit (Apple Silicon)
GOOS=darwin GOARCH=arm64 go build -o openCursor-macos-arm
# Linux 64-bit
GOOS=linux GOARCH=amd64 go build -o openCursor-linux
# Linux ARM64
GOOS=linux GOARCH=arm64 go build -o openCursor-linux-arm64The project includes a Makefile for common tasks:
# Build the project
make build
# Build for all platforms
make build-all
# Clean build artifacts
make clean
# Run tests
make test
# Install to local bin
make installAfter building, verify the binary works correctly:
# Check version and help
./openCursor --help
# Test with a simple query (requires API key)
export OPENAI_API_KEY="your-api-key"
./openCursor "Hello, test message"For development work:
# Install development dependencies
go mod download
# Run with go run (for testing)
go run main.go "test query"
# Run tests
go test ./...
# Format code
go fmt ./...
# Check for issues
go vet ./...openCursor/
├── cmd/ # Command line interface
├── internal/ # Internal packages
│ ├── client/ # AI client implementation
│ └── tools/ # Tool management
├── main.go # Application entry point
├── go.mod # Go module definition
├── go.sum # Go module checksums
├── Makefile # Build automation
└── README.md # Project documentation
We welcome contributions! Please feel free to submit issues and pull requests.
MIT License
一个开源的 openCursor 项目，主要实现了 Cursor Agent 的核心流程，方便大家学习和使用。本项目展示了 Cursor Agent 功能背后的基础工作流程。
🎯 特别说明: 本项目包括此 README.md 文件在内，完全由 Cursor 自动生成。程序员仅使用了对话面板，没有在编辑器中编写一行代码。这展示了AI驱动开发的强大能力，类似于 Claude Code 等标杆性CLI工具可能成为一种新的轻量级AI Coding交互方式。
🔧 技术细节: 本代码库使用的提示词和工具都是复刻自 Cursor 官方原生工具，提供了对原始 Cursor Agent 工作流程的真实还原。
- 🚀 开源实现: 完整实现 Cursor Agent 核心工作流程
- 🤖 AI 集成: 默认集成 DeepSeek AI 模型
- 🔧 简单配置: 简单的环境变量设置
- 🌐 API 兼容: 支持 OpenAI API 兼容的模型
- 📦 便携性: 单一二进制可执行文件
- 💡 教育性: 非常适合学习 AI 代理工作流程
看看 openCursor 的实际效果！以下展示如何用一条命令创建完整的贪吃蛇游戏：
openCursor "写一个贪吃蛇游戏，使用HTML和JavaScript。包含完整的游戏控制、计分系统和游戏结束功能。让它具有响应式设计且视觉效果良好。"这个命令将生成一个完整的、可玩的贪吃蛇游戏，包括：
- ✅ 带画布的HTML结构
- ✅ JavaScript游戏逻辑
- ✅ 键盘控制（WASD或方向键）
- ✅ 分数追踪
- ✅ 游戏结束检测
- ✅ 响应式设计
- ✅ 现代化样式
AI代理将创建所有必要的文件，为你提供一个可以立即在浏览器中运行的完全功能的游戏！
下载 openCursor 二进制文件并放到系统的 bin 目录：
# 下载二进制文件（请替换为实际下载地址）
wget https://github.com/your-username/openCursor/releases/latest/download/openCursor
# 添加执行权限
chmod +x openCursor
# 移动到系统 bin 目录
sudo cp openCursor /usr/local/bin/
# 现在可以在任意目录使用
openCursor --help方式1：环境变量配置
如果你使用 zsh，添加到 ~/.zshrc 文件：
export OPENAI_API_KEY="你的-deepseek-api-密钥"如果你使用 bash，添加到 ~/.bashrc 文件：
export OPENAI_API_KEY="你的-deepseek-api-密钥"方式2：自定义模型配置
你也可以配置其他 OpenAI API 兼容格式的模型：
export OPENAI_API_KEY="你的-api-密钥"
export MODEL="deepseek-chat"
export BASE_URL="https://api.deepseek.com/v1"# 基本用法
openCursor "你好，请介绍一下你自己"
# 编程助手
openCursor "写一个计算斐波那契数列的Python函数"
# 代码审查
openCursor "帮我审查这段代码，看看有什么改进建议"- 访问 DeepSeek 开放平台
- 注册并登录账户
- 前往 API 密钥管理页面
- 创建新的 API 密钥
- 复制密钥并配置到环境变量中
📖 详细的API文档和使用示例，请访问: DeepSeek API 文档
- Go 1.19+: 确保安装了 Go 1.19 或更高版本
- Git: 用于克隆代码库
- Make: 用于使用 Makefile（可选）
检查你的 Go 版本：
go version# 克隆代码库
git clone https://github.com/your-username/openCursor.git
cd openCursor
# 下载依赖
go mod download
# 构建项目
go build -o openCursor
# 或者使用 Make
make build标准构建：
go build -o openCursor优化构建（更小的二进制文件）：
go build -ldflags="-s -w" -o openCursor调试构建：
go build -gcflags="all=-N -l" -o openCursor-debug为不同的操作系统和架构构建：
# Windows 64位
GOOS=windows GOARCH=amd64 go build -o openCursor.exe
# macOS 64位 (Intel)
GOOS=darwin GOARCH=amd64 go build -o openCursor-macos-intel
# macOS 64位 (Apple Silicon)
GOOS=darwin GOARCH=arm64 go build -o openCursor-macos-arm
# Linux 64位
GOOS=linux GOARCH=amd64 go build -o openCursor-linux
# Linux ARM64
GOOS=linux GOARCH=arm64 go build -o openCursor-linux-arm64项目包含一个 Makefile 用于常见任务：
# 构建项目
make build
# 为所有平台构建
make build-all
# 清理构建产物
make clean
# 运行测试
make test
# 安装到本地 bin 目录
make install构建完成后，验证二进制文件是否正常工作：
# 检查版本和帮助信息
./openCursor --help
# 使用简单查询测试（需要 API 密钥）
export OPENAI_API_KEY="你的-api-密钥"
./openCursor "你好，测试消息"开发工作的环境设置：
# 安装开发依赖
go mod download
# 使用 go run 运行（用于测试）
go run main.go "测试查询"
# 运行测试
go test ./...
# 格式化代码
go fmt ./...
# 检查问题
go vet ./...openCursor/
├── cmd/ # 命令行界面
├── internal/ # 内部包
│ ├── client/ # AI 客户端实现
│ └── tools/ # 工具管理
├── main.go # 应用程序入口
├── go.mod # Go 模块定义
├── go.sum # Go 模块校验和
├── Makefile # 构建自动化
└── README.md # 项目文档
欢迎贡献代码！请随时提交 issue 和 pull request。
MIT 许可证