About this project
Supermemory is a memory and context engine for AI applications, designed to give AI assistants persistent memory across conversations. It automatically extracts facts from conversations, builds user profiles, handles knowledge updates and contradictions, and forgets expired information. The project provides a single API covering memory, RAG, user profiles, connectors (Google Drive, Gmail, Notion, OneDrive, GitHub), and multi-modal file processing (PDFs, images, videos, code). It offers plugins for popular AI tools like Claude Code, Cursor, Codex, OpenCode, and an MCP server. Developers can integrate via npm or pip packages with framework wrappers for Vercel AI SDK, LangChain, LangGraph, OpenAI Agents SDK, Mastra, and others. For self-hosting, a single binary runs locally with zero config, supporting any model or fully offline via Ollama. The project claims state-of-the-art results on memory benchmarks (LongMemEval, LoCoMo, ConvoMem) and provides an open-source benchmark framework (MemoryBench).
Comments
0 Rating appears after 10 ratings
Sign in to join the discussion.