Named after the Egyptian god of wisdom, writing, and magic

Natural language to shell commands, powered by local LLMs.

Describe what you want in plain English, and Thoth translates it to the right shell command for your OS.

$ thoth find all files larger than 100mb
$ find . -size +100M -type f
./Downloads/movie.mp4
./Documents/backup.zip

• Natural language input - No need to remember command syntax
• OS-aware - Generates macOS, Ubuntu, Fedora, Arch, or Windows commands automatically
• Local & private - Uses Ollama, everything runs on your machine
• Zero config - Just install and use
• Fast - Direct execution, no confirmation prompts

# macOS
brew install ollama

# Linux
curl -fsSL https://ollama.com/install.sh | sh

ollama pull gemma3

From releases (recommended):

# macOS (Apple Silicon)
curl -L https://github.com/HopperShell/thoth/releases/latest/download/thoth-macos-arm64 -o /usr/local/bin/thoth
chmod +x /usr/local/bin/thoth

# macOS (Intel)
curl -L https://github.com/HopperShell/thoth/releases/latest/download/thoth-macos-x64 -o /usr/local/bin/thoth
chmod +x /usr/local/bin/thoth

# Linux (x64)
curl -L https://github.com/HopperShell/thoth/releases/latest/download/thoth-linux-x64 -o /usr/local/bin/thoth
chmod +x /usr/local/bin/thoth

From source:

git clone https://github.com/HopperShell/thoth
cd thoth
cargo build --release
cp target/release/thoth /usr/local/bin/

thoth config

Select your model from your installed Ollama models:

? Select your model:
> gemma3:latest (2.3 GB)
  llama3:latest (4.7 GB)
  mistral:latest (4.1 GB)

? Ollama URL [http://localhost:11434]:

✓ Saved to ~/.config/thoth/config.toml

thoth show disk usage
thoth find what is using port 3000
thoth list all running docker containers
thoth compress this folder into a tar.gz

Run thoth config to configure your model and Ollama URL interactively.

Config location:

• macOS: ~/Library/Application Support/thoth/config.toml
• Linux: ~/.config/thoth/config.toml

model = "gemma3:latest"
ollama_url = "http://localhost:11434"

To change models, just run thoth config again.

1. Thoth detects your OS (macOS, Ubuntu, Fedora, Arch, etc.)
2. Sends your query + OS context to a local Ollama model
3. Parses the generated command
4. Executes it directly in your shell

All processing happens locally. Nothing is sent to external servers.

• Ollama running locally
• A pulled model (gemma3, llama3, mistral, etc.)

MIT