Provides AI agents with a searchable, long-term memory infrastructure by indexing markdown notes, documents, and code.
Sponsored
Empower AI agents with a robust long-term memory system by transforming your markdown notes, documents, and code into a highly searchable knowledge base. Memtomem indexes plain text files, enabling AI agents to query information using both keyword (BM25) and semantic search. This prevents AI agents from forgetting critical context across sessions and ensures relevant content is always accessible, consolidating scattered notes into a single, vendor-agnostic index where your files remain the source of truth.
主な機能
010 GitHub stars
02Hybrid search combining BM25 keyword, dense vector, and RRF fusion
03Semantic chunking for various file types including Markdown, Python, JS/TS, JSON/YAML/TOML
04Incremental indexing with chunk-level SHA-256 diffs for efficiency
05Namespaces for organizing memories with auto-derivation from folder names
06Built-in memory maintenance features like near-duplicate detection, time-based decay, and auto-tagging
ユースケース
01Providing AI agents with persistent, searchable knowledge across multiple sessions
02Allowing AI agents to retrieve specific information from personal or project knowledge bases using keywords and semantic meaning
03Organizing and indexing diverse project files (markdown, code, data) into a unified, searchable memory for AI agents
