Obsidian-backed knowledge graph with semantic search, entity extraction, and cross-session memory. 11 MCP tools. Works with Claude Code, Cursor, Windsurf, and any MCP-compatible editor.
Open, local-first engram memory for AI agents: read, correct, and delete what your agent remembers. Works across Claude Code, Hermes, OpenClaw, and Cursor over MCP.
Exposes a shared graph-based context engine as MCP tools for Claude, Copilot, and other AI agents, enabling knowledge ingestion, recall, search, and LLM-ready context assembly across sessions.
Exposes search, locate, expand, and find_related tools over MCP so AI coding agents in Claude Code, Cursor, Copilot, and other clients can retrieve the best code chunks (path, lines, snippet) by meaning instead of grep or glob. This hybrid code search lets agents find the right files in fewer steps, using up to 50% fewer tokens.