hi-shell 🐚

👋 hi-shell: An intelligent terminal assistant that translates your natural language descriptions into executable bash commands.

hi-shell helps you bridge the gap between "what I want to do" and "how do I write that command?". Whether you're a terminal veteran or a newcomer, hi-shell provides a fast, AI-powered way to generate and execute commands safely.

Should be test before 1.0.0 release:

• embedded models
• local models (tested lmstudio and ollama)
• cloud models (tested openrouter, gemini, custom, need to test with antrophic)
• interactive mode
• one-shot mode
• safety features
• telemetry
• release workflow works
• cargo publish works
• homebrew update works
• scoop update works
• works on linux
• works on windows

✨ Features

• Multi-LLM Support:
  • Embedded: Run models locally using candle with hardware acceleration (Metal/CUDA). Supports Llama, Phi-3, and Qwen2 architectures.
  • Local: Connect to your own Ollama or LM Studio instance.
  • Cloud: Integration with OpenRouter, Gemini, and Anthropic.
• Interactive REPL: A dedicated shell environment for continuous assistance.
• One-shot Mode: Get quick answers directly from your command line.
• Safety First: Dangerous commands are flagged, and confirmation is required before execution.
• Telemetry: Optional anonymous usage stats to help improve the tool.

🚀 Installation

You can install hi-shell using your preferred method:

⚡ Quick Install (macOS & Linux)

curl -sSL https://raw.githubusercontent.com/tufantunc/hi-shell/main/install.sh | bash

🍏 Homebrew (macOS & Linux)

brew tap tufantunc/tap
brew install hi-shell

🪟 Scoop (Windows)

scoop bucket add hi-shell https://github.com/tufantunc/scoop-bucket
scoop install hi-shell

🦀 Cargo

If you have Rust installed, you can install it via Cargo:

# Install via crates.io
cargo install hi-shell

# OR install pre-compiled binary via cargo-binstall
cargo binstall hi-shell

📦 Manual Download

You can download the pre-built binaries for your operating system from the Releases page.

1. Download the version corresponding to your OS (macOS, Linux, or Windows).
2. Move the binary to a folder in your PATH.
3. Run hi-shell --init to set up your preferred LLM provider.

🔄 Updating

🍏 Homebrew (macOS & Linux)

brew update && brew upgrade hi-shell

🪟 Scoop (Windows)

scoop update hi-shell

🦀 Cargo

cargo install hi-shell
# OR
cargo binstall hi-shell

⚡ Quick Install Script

Re-run the install script to get the latest version:

curl -sSL https://raw.githubusercontent.com/tufantunc/hi-shell/main/install.sh | bash

🛠 Development

If you'd like to contribute or build from source, follow these steps.

Prerequisites

• Rust (latest stable version recommended)
• C Compiler (for some dependencies)

Cloning the Repository

git clone https://github.com/tufantunc/hi-shell.git
cd hi-shell

Building from Source

You can build the project for your specific platform using Cargo.

🍏 macOS (with Metal support)

cargo build --release --features metal

🐧 Linux

cargo build --release

𝝿 Linux (with CUDA support)

cargo build --release --features cuda

🪟 Windows

cargo build --release

The compiled binary will be located in target/release/hi-shell.

🧠 Embedded Models

When using the embedded LLM provider, hi-shell downloads and runs GGUF models locally.

Supported Architectures

• Llama (Llama 2, Llama 3, Llama 3.2)
• Phi-3 (Phi-3-mini)
• Qwen2 (Qwen2, Qwen2.5, Qwen2.5-Coder)

Recommended Quantization

Use Q4_K_M or Q5_K_M quantization formats for best compatibility. Lower quantization levels (Q2_K) may not be supported.

Model Management

# View and manage downloaded models
hi-shell --models

🤝 Contributing

Contributions are welcome! Please read our Contributing Guide before submitting issues or pull requests.

📄 License

This project is licensed under the MIT License.