
Lahfir
172 posts

Lahfir
@mdlahfir
Computer Use | Built agent-desktop, cracked-agent, PILOT | 2X founder




Jarvis is finally here! Happy to launch agent-desktop skill for @openclaw Now your claw can: Click, Drag, Open, Access every single app in your Mac (+54 commands). Windows and Linux coming soon! This is a fully headless mode achieved using accessibility trees. Works on Electron apps as well (Slack, Discord, etc.) It's fully Open Source. Excited to see what you all will build with it! Inspired by agent-browser! Thanks @ctatedev


Jarvis is finally here! Happy to launch agent-desktop skill for @openclaw Now your claw can: Click, Drag, Open, Access every single app in your Mac (+54 commands). Windows and Linux coming soon! This is a fully headless mode achieved using accessibility trees. Works on Electron apps as well (Slack, Discord, etc.) It's fully Open Source. Excited to see what you all will build with it! Inspired by agent-browser! Thanks @ctatedev




Here's the new Clicky. It's the simplest interface in the world to talk to AI + spawn agents. It builds Mac apps. It does research to help you find IG micro-influencers. It interacts with native Apple Notes, Calendar, Reminders. Built for consumers, 0 setup. Try today, free.

Jarvis is finally here! Happy to launch agent-desktop skill for @openclaw Now your claw can: Click, Drag, Open, Access every single app in your Mac (+54 commands). Windows and Linux coming soon! This is a fully headless mode achieved using accessibility trees. Works on Electron apps as well (Slack, Discord, etc.) It's fully Open Source. Excited to see what you all will build with it! Inspired by agent-browser! Thanks @ctatedev




hear me out: what if Clicky can actually Click? I have crawled through the depths of hell to bring you, for the foreseeable years, one of the more important foundational pieces of computer use (certainly in implementation): Fast, accurate, and comprehensive accessibility tree interaction in pure Rust, usable for laying out entire desktop apps without needing the LLMs to guess the button co-ordinates from screenshots! built using agent-desktop, an extremely fast Rust-based accessibility tree retrieval and interaction framework for Mac (GA), Windows, and Linux Made clicky more powerful by making Claude make tool calls using the agent-desktop CLI to retrieve the accessibility tree of the app instead of the screenshot to actually access the interactive elements of the app, instead of burning tokens by passing the screenshots on every iteration. Here's how it works: 1. Claude just receives context about the app in the frontmost part of the screen 2. A custom tool that can only execute agent-desktop CLI commands 3. Claude can use this tool to execute 17+ powerful agent-desktop commands to retrieve and interact with the app. 4. In show-only mode, it just shows what and where, but in interact mode, it uses Agent-Desktop to interact with the app.














