macOS agent that turns prompts into automated clicks - Openclick

openclick is an experimental open-source CLI that drives macOS UI from a prompt. An LLM produces a plan of UI actions; openclick executes them via the macOS Accessibility APIs. Early and rough. MIT-licensed.

Framer — Launch websites with enterprise needs at startup speeds.

Launch websites with enterprise needs at startup speeds.

hey y'all! Excited to share a cool tool I've been building: OpenClick 🖱️ OpenClick is an open-source macOS automation system that turns natural-language prompts into real desktop actions. It's very simple, just type: openclick run "open Calculator and calculate 17 times 23" Under the hood: a signed local helper drives the Mac via Accessibility + Screen Recording, while a planner/executor loop uses screenshots, AX trees, and verification to keep working through multi-step tasks. It learns from each run via local memories. Screenshots and context go directly to your chosen model provider (Anthropic, OpenAI, or local) - there's no OpenClick backend in the middle. Works today for browser navigation, Gmail, Finder, Calculator, TextEdit-style flows, and common shortcuts. Still early (0.2.0 beta), still rough on canvas-heavy apps, but it's a real glimpse of where local desktop agents are heading. PS: It feels weird to be back on Product Hunt after such a long time! I missed it!

Riccardo congrats on the launch. screenshot-driven planners are interesting territory. one thing we keep getting tripped up on in browser automation: the screenshot the LLM saw at step N is stale by step N+1 if the page re-rendered. how does Openclick handle that. does it re-anchor each step from a fresh AX tree, or trust the original plan and verify at the end? curious because the AX-tree-vs-pixel-coords delta is exactly where most of our agent failures hide.

Framer — Launch websites with enterprise needs at startup speeds.

Launch websites with enterprise needs at startup speeds.