dnakov

dnakov / computer-use

Public

macOS computer use CLI — screenshots, input simulation, app management, session orchestration

15
1
100% credibility
Found Mar 31, 2026 at 15 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Swift
AI Summary

computer-use is a macOS command-line tool for desktop control, including screenshots, mouse and keyboard input, app and window management, with optional sessions for added safety guards.

Star Growth

See how this repo grew from 15 to 15 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is computer-use?

This Swift CLI turns your macOS desktop into a programmable interface, letting you capture screenshots, simulate mouse/keyboard input, manage apps and windows, and orchestrate sessions with safety guards. Built for macOS 14+, it outputs JSON for easy piping into scripts or LLMs, solving the pain of brittle AppleScript or UI automation for computer use agents. Install via Homebrew and grant Accessibility/Screen Recording perms—raw commands like `computer-use input type-text "hello"` or session-based actions like `computer-use session action left_click --coordinate 500,300`.

Why is it gaining traction?

As an open source computer use GitHub repo inspired by Anthropic's computer-use preview, it stands out with session tiers (full/click/read access), auto app-hiding for clean screenshots, pixel-to-screen coord conversion, and clipboard protection—features proprietary tools like computer-use-mcp charge for. The teach mode overlays native tooltips for step-by-step demos, and batch actions handle complex flows without race conditions. Devs dig the native binary speed over Python wrappers.

Who should use this?

AI engineers prototyping computer-use agents or computer-use GitHub Anthropic integrations on macOS. Automation hackers replacing flaky pyautogui with precise input sim and app grants. No-go for Windows/Linux or production without more battle-testing.

Verdict

Grab it for macOS experiments—294 tests and Homebrew formula show polish despite 15 stars and 1.0% credibility score. Early but promising open source computer use tool; fork and contribute if you need computer-use-preview Gemini/OpenAI parity.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.