Local AI Worker is a desktop app (Tauri + Rust) that runs autonomous workers against
Ollama in Docker: scheduled tasks, domain guardrails, optional repo-backed agents with guarded
git/gh, and secrets stored in your OS keychain.
What you can do
-
Run Ollama inside Docker via bundled Compose — the app starts the stack automatically when you enable workers that use a loopback Ollama tier (
localhost/127.0.0.1), or you can bring it up manually from diagnostics. Pull models inside the container when prompted in the docs. - Define workers with maintenance domains, escalation tiers (Ollama and optional Cursor), and cadence or one-shot tasks.
- Optionally attach a GitHub repository for the structured JSON repo agent loop (
observe/apply_git/apply_github). - Use audit hooks and rolling guardrails without leaving your machine for inference.
Start here
- Getting started — install, Ollama stack, token, first worker.
- User guide — full UI tour, flows, workers, repo agent, data locations.
- Downloads — latest installers from GitHub Releases (linked by this site).
- Architecture — how the UI, host, core library, Docker, and Ollama connect.
Overview
Environment panel in the desktop app (hardware hints, Ollama stack, quick start).
