mcp-vnc and claude-kvm
These are competitors offering similar VNC-based remote desktop control capabilities for macOS, with hrrrsn/mcp-vnc appearing to be the more established project based on community adoption (29 vs 2 stars).
About mcp-vnc
hrrrsn/mcp-vnc
Remote desktop control for AI
Implements the Model Context Protocol (MCP) to expose VNC remote desktop control as tools for AI agents, supporting mouse clicks, keyboard input, text typing, and screenshot capture across Windows, Linux, and macOS. Integrates via stdio transport with Claude Desktop and VS Code's Copilot, connecting to standard VNC servers through configurable host/port/password credentials.
About claude-kvm
ARAS-Workspace/claude-kvm
🤖 ⚡️ MCP server ( MacOS) — control remote desktops via VNC
Implements a two-layer architecture: a Node.js MCP proxy communicating with Claude over stdio JSON-RPC, and a native Swift/C daemon handling VNC connections with coordinate scaling, differential screen capture, and on-device OCR via Apple Vision framework. Enables Claude to control remote desktops through precise mouse/keyboard input, real-time frame capture with progressive verification strategies, and automatic resolution scaling to optimize token efficiency.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work