Loopi — desktop automation with AI agents

★ 196

Loopi is a local desktop automation platform: a visual workflow builder, control of a real Chromium browser, cursor and keyboard automation, shell command execution, AI agents and 80+ built-in integrations. Reach for it when you are tired of moving data between services by hand, when you need to automate a program that has no API by driving the mouse and keyboard, when you want an agent that opens a browser and fills in forms itself, or when you want the tool itself to run on your own machine. Inside: a ReactFlow editor, browser automation (navigation, clicks, screenshots, data extraction), desktop control (Linux via ydotool on Wayland and nut-js on X11, plus macOS and Windows), shell commands that capture stdout, stderr and exit codes into variables, typed variables, scheduling, a credentials manager and a CLI. Goal-driven agents on a schedule judge each run and can patch a broken workflow; models can be OpenAI, Anthropic, the Claude Code CLI or local Ollama. Installers are an .exe for Windows and a .deb for Linux, or build from source with pnpm. Data stays local only when you use the local Ollama model: cloud models and services receive whatever you send them. The license is not declared in the repository metadata.