Side-by-side comparison of two Model Context Protocol servers — install paths, tools, quality signals, and directory engagement so you can pick the right one for Claude, Cursor, and other MCP clients.
Local-first desktop context server for AI agents. Captures screen (OCR), voice (Whisper), keyboard/mouse activity, and clipboard. Exposes 35 MCP tools for screen capture, voice transcription, activity history, semantic memory, and project detection. Zero cloud dependency. AGPL-3.0.
TypeScript MCP server for Kanboard with multi-project routing via .kanboard.yaml. 37 tools, 982 unit tests, Docker multi-arch on GHCR, provenance-signed npm releases via OIDC, Diátaxis-structured docs, MIT licensed.