GhostDesk
Active·★ 155·NOASSERTION·Updated 2026-09-23
GhostDesk is an MCP server that gives LLM agents a sandboxed Linux desktop with vision and input tools. It allows agents to interact with any GUI application, perform multi-step workflows, and work autonomously.
#agentic#ai-agent#automation#autonomous-agent#browser-automation#claude#computer-use#docker
01
Features
01Vision and input tools: screenshot, click, type, scroll, etc.
02Sandboxed desktop in Docker with full isolation
03Works with any MCP-compatible LLM (Claude, GPT, Gemini)
04Operates any GUI application without API integration
05Human-like input simulation (Bézier curves, variable delays)
02
Compatibility
Linux
Linux
Verified via docs
Docker
Docker
Verified via docs
03
Quick start
1
$ docker run -d --name ghostdesk-my-agent --restart unless-stopped -p 3000:3000 -p 5900:5900 -p 6080:6080 -v ghostdesk-my-agent-home:/home/agent --shm-size 2g -e VNC_PASSWORD=changeme -e TZ=UTC -e LOCALE=en_US.utf8 ghcr.io/yv17labs/ghostdesk:latest
04
Use cases
↳Web scraping and data processing (e.g., Amazon to Google Sheets)
↳Flight search and price comparison across multiple sites
↳Operating legacy or internal software without API access
05
Alternatives
GenericAgent★ 14.3k
Self-evolving agent: grows skill tree from 3.3K-line seed, achieving full system control with 6x less token consumption
camofox-mcp★ 119
Anti-detection browser MCP server for AI agents — navigate, interact, and automate the web without getting blocked
cdpilot★ 32
Zero-dependency browser automation CLI. 70+ commands, 10 test assertions, smart commands (click/fill by text — no LLM needed). MCP server for AI agents with 500x fewer tokens. Extract, observe, script runner. 50KB, pure CDP.
open-claude-in-chrome★ 255
Claude in Chrome, reverse-engineered and open-source. No domain blocklist. Any Chromium browser. Same 18 MCP tools, same performance.
Comments
Log in to leave a comment
No comments yet. Be the first!