Setup

Prerequisites

• Supported OS — Use a package built for your operating system and architecture (available from the download page).
• AI backend — A local LLM server (for example LLaMA.cpp or Ollama) or an external OpenAI-compatible API.

Install from a prebuilt package

1. Download the package for your platform. The official site serves olaf.tar.gz at this link; other formats may come from your organization’s mirror.

2. Extract the tarball (or run your installer), then locate the olaf binary:

   tar -xzf olaf.tar.gz
   

3. Make it executable (Unix-like systems), if needed:

   chmod +x olaf
   

4. Put it on your PATH — Either move the binary to a directory already on your PATH, or add its directory:

   export PATH="/path/to/olaf/dir:$PATH"
   

   For a permanent setup, use your shell profile or system-wide path configuration.

5. Verify:

   olaf --help
   

   (Exact flags depend on the shipped binary; --help or running without arguments may print usage.)

Run

After configuring backends (see Configuration):

olaf

Use your platform’s terminal or SSH session; Olaf is a full-screen TUI.

Gateway (optional)

Some deployments include olaf-gateway as a separate binary in the same package. See your release notes or administrator documentation for proxy setup and the optional web UI served at /web/ on the gateway port.