Memoriki combines LLM-powered wikis with semantic search and knowledge graphs to automate personal knowledge base maintenance across any MCP-compatible AI agent.

Merges Karpathy's LLM Wiki pattern with MemPalace MCP server to create self-maintaining structured markdown wikis with [[wiki-links]] and YAML metadata

Uses ChromaDB for semantic search by meaning rather than keywords, plus a temporal knowledge graph that tracks relationship changes over time

Differs from RAG by compiling knowledge once into persistent wiki pages that stay current, rather than re-deriving answers from raw chunks each query

Works with any MCP-compatible agent including Claude Code, OpenAI Codex, Cursor, and Gemini CLI