Quickstart
Huginn works with any model backend — managed (built-in runtime, no external deps), Ollama, or cloud providers like Anthropic and OpenRouter. For a fully self-contained setup, see managed mode.
Install Huginn
curl -fsSL https://huginn.sh/install.sh | sh
Or with Homebrew on macOS:
brew install huginn
Run it
Navigate to any git repository and run:
cd ~/projects/my-app
huginn
First run
Huginn indexes your repository on startup and drops you into the TUI:
huginn v0.2.0 ready · 847 files indexed
you /plan add error handling to the authentication flow
huginn Planning with qwen3-coder:30b...
Plan:
1. Wrap handler in internal/auth/handler.go with error middleware
2. Add structured error types in internal/auth/errors.go
3. Update tests
❯ [a]pprove [e]dit [c]ancel
Prerequisites
No external dependencies required in managed mode. Choose your backend:
- Managed (default for new installs): Huginn downloads its own runtime — nothing else needed
- Ollama: Install Ollama and pull a model:
ollama pull qwen2.5-coder:14b - Cloud API key: Set provider + key in
~/.huginn/config.json(see Configuration)
Your first request
/plan add error handling to the authentication flow
Huginn will analyze your codebase, generate a structured plan, ask you to approve, then apply the changes.
Managing models
# See what's installed
huginn models
# Browse available models
huginn models --available
# Download a different model
huginn pull qwen2.5-coder:7b
Next steps
- Named Agents — create persistent AI personas
- Configuration — customize model slots and settings
- Commands — all slash commands