A production-ready, model-agnostic CLI coding agent with safety-first design
clippy-code is an AI-powered development assistant that works with any OpenAI-compatible API provider. It features robust permission controls, streaming responses, and multiple interface modes for different workflows.
# Install uv if you haven't already
curl -LsSf https://astral.sh/uv/install.sh | sh
# Run clippy-code directly - no installation needed!
uvx clippy-code "create a hello world python script"
# Start interactive mode
uvx clippy-codeclippy-code supports multiple LLM providers through OpenAI-compatible APIs:
# OpenAI (default)
echo "OPENAI_API_KEY=your_api_key_here" > .env
# Or choose from many supported providers:
echo "ANTHROPIC_API_KEY=your_api_key_here" > .env
echo "MISTRAL_API_KEY=your_api_key_here" > .env
echo "CEREBRAS_API_KEY=your_api_key_here" > .env
echo "GROQ_API_KEY=your_api_key_here" > .env# One-shot mode - execute a single task
clippy "create a hello world python script"
# Interactive mode - REPL-style conversations
clippy
# Specify a model
clippy --model gpt-4 "refactor main.py to use async/await"
# Auto-approve all actions (use with caution!)
clippy -y "write unit tests for utils.py"clippy-code can dynamically discover and use tools from MCP (Model Context Protocol) servers. For detailed configuration and available servers, see docs/MCP.md.
Quick setup:
# Create the clippy directory
mkdir -p ~/.clippy
# Copy the example configuration
cp mcp.example.json ~/.clippy/mcp.json
# Edit it with your API keys- 🌐 Broad Provider Support: OpenAI, Anthropic, Cerebras, Groq, Mistral, Ollama, and many more
- 🛡️ Safety-First Design: Three-tier permissions with interactive approval for risky operations
- 🔄 Multiple Interface Modes: One-shot tasks, interactive REPL, and rich document mode
- 🤖 Advanced Agent Capabilities: Streaming responses, context management, subagent delegation
- 🔌 Extensible Tool System: Built-in file operations, command execution, and MCP integration
- 💻 Developer Experience: Type-safe codebase, rich CLI, flexible configuration
clippy-code provides smart file operations with validation for many file types:
| Tool | Description | Auto-Approved |
|---|---|---|
read_file |
Read file contents | ✅ |
write_file |
Write files with syntax validation | ❌ |
delete_file |
Delete files | ❌ |
list_directory |
List directory contents | ✅ |
create_directory |
Create directories | ❌ |
execute_command |
Run shell commands | ❌ |
search_files |
Search with glob patterns | ✅ |
get_file_info |
Get file metadata | ✅ |
read_files |
Read multiple files at once | ✅ |
grep |
Search patterns in files | ✅ |
edit_file |
Edit files by line (insert/replace/delete/append) | ❌ |
write_file includes syntax validation for Python, JSON, YAML, HTML, CSS, JavaScript, TypeScript, Markdown, Dockerfile, and XML.
clippy-code works with any OpenAI-compatible provider: Anthropic, Chutes, Cerebras, Gemini, Groq, LM Studio, MiniMax, Mistral, Ollama, OpenAI, OpenRouter, Synthetic.new, Together AI, Z.AI, and more.
# List available providers
/model list
# Try a model without saving
/model use cerebras qwen-3-coder-480b
# Save a model configuration
/model add cerebras qwen-3-coder-480b --name "q3c"
# Switch to a saved model
/model q3c- Provider-specific API keys:
ANTHROPIC_API_KEY,CHUTES_API_KEY,CEREBRAS_API_KEY,GOOGLE_API_KEY,GROQ_API_KEY,MISTRAL_API_KEY,OPENAI_API_KEY, etc. OPENAI_BASE_URL- Optional base URL override for custom providers
# Clone and enter repository
git clone https://github.com/yourusername/clippy.git
cd clippy
# Create virtual environment with uv
uv venv
source .venv/bin/activate # or .venv\Scripts\activate on Windows
# Install in editable mode with dev dependencies
uv pip install -e ".[dev]"
# Run clippy in development
python -m clippy
# For normal usage, use uvx clippy-code# Format code
make format
# Run linting, type checking, and tests
make check
make testFor detailed information about:
- Adding new tools: See CONTRIBUTING.md
- MCP server integration: See docs/MCP_DOCUMENTATION.md
- Subagent development: See docs/SUBAGENTS.md
- OpenAI Compatibility: Single standard API format works with any OpenAI-compatible provider
- Safety First: Three-tier permission system with user approval workflows
- Type Safety: Fully typed Python codebase with MyPy checking
- Clean Code: SOLID principles, modular design, Google-style docstrings
- Streaming Responses: Real-time output for immediate feedback
- Quick Start Guide - Getting started in 5 minutes
- MCP Integration - Model Context Protocol setup and usage
- Contributing Guide - Development workflow and code standards
- Agent Documentation - Internal architecture for developers
Made with ❤️ by the clippy-code team