manushi4/Screenhand

Give AI eyes and hands on your desktop. Open-source MCP server for desktop automation — screenshots, UI control, browser automation, OCR. Works with Claude, Cursor, and any MCP client. macOS + Windows.

46
/ 100
Emerging

Implements native OS Accessibility APIs and Chrome DevTools Protocol for ~50ms action latency without requiring LLM calls per interaction, complemented by a persistent "App Mastery Map" that learns spatial UI blueprints and navigation patterns through normal tool usage. Includes 111 tools spanning desktop control, browser automation, smart fallbacks (Accessibility → CDP → OCR → coordinates), memory/learning systems, and multi-agent job orchestration, with prebuilt knowledge for 36+ applications and 49 automation playbooks loaded automatically on app detection.

Available on npm.

Maintenance 13 / 25
Adoption 6 / 25
Maturity 18 / 25
Community 9 / 25

How are scores calculated?

Stars

16

Forks

2

Language

TypeScript

License

AGPL-3.0

Last pushed

Mar 11, 2026

Commits (30d)

0

Dependencies

4

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/mcp/manushi4/Screenhand"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.