Autohand Code CLI
Autohand Code CLI is a fast, terminal-native AI coding agent that plans, edits, tests, and automates work across your codebase. It solves the problem of context switching by letting developers describe outcomes in natural language and execute approved changes directly from the CLI. It is for developers who want an AI assistant integrated into their terminal workflow.
✨ Key features
- Terminal-native agent with planning and tool execution
- Interactive REPL with file mentions, slash commands, and shortcuts
- Modular skills for specialized workflows
- Multi-provider support: OpenRouter, OpenAI, Bedrock, local models
- Git integration with automatic commits
- Cross-platform: macOS, Linux, Windows
🎯 Use cases
- Refactor code with natural language prompts
- Add features like dark mode toggles
- Fix TypeScript errors with auto-confirmation
- Automate repetitive engineering tasks in CI/CD
- Generate git patches without applying changes
📦 Installation
🧰 Requirements: Requires Bun ≥1.0, Git for version control, and optionally ripgrep for faster search. Works on macOS, Linux, and Windows.
Quick Install (Recommended)
curl -fsSL https://autohand.ai/install.sh | bash
Homebrew
brew install autohandai/code/autohand-code
Manual Installation
git clone https://github.com/autohandai/code-cli.git
cd code-cli
bun install
bun run buildbun add -g .🚀 Usage
autohandautohand -p "add a dark mode toggle to the settings page"autohand -p "fix the TypeScript errors" -yautohand -p "refactor the auth module" -c❓ FAQ
How do I install Autohand Code CLI?
You can install via the quick install script (curl -fsSL https://autohand.ai/install.sh | bash), Homebrew (brew install autohandai/code/autohand-code), or manually by cloning and building with Bun.
What are the system requirements?
You need Bun ≥1.0, Git for version control features, and optionally ripgrep for faster search. It works on macOS, Linux, and Windows.
Can I use it in non-interactive mode?
Yes, you can use command mode with the -p or --prompt flag to run a single instruction, and options like -y for auto-confirmation and -c for auto-commit.
Does it support multiple LLM providers?
Yes, it supports OpenRouter, LLMGateway, OpenAI, AWS Bedrock, DeepSeek, Azure Foundry Models, Z.ai, and local models.
📊 Repository
🤖 Overview, features, install steps and FAQ were generated from the project's README on Sep 4, 2026. Always check the original source before running commands.