TL;DR
Supermemory is a memory engine that provides fact extraction, user profiles, hybrid search, and external service integration via a single API, enabling AI agents to maintain context across conversations.
Key features
Memory: Extracts facts from conversations, handles temporal changes, contradictions, and automatic forgetting
User profiles: Automatically maintains stable facts and recent activity, ~50ms response
Hybrid search: Integrates RAG and memory in a single query
Connectors: Auto-sync and webhook support for Google Drive, Gmail, Notion, OneDrive, GitHub, etc.
Multimodal extractor: Processes PDFs, images (OCR), videos (subtitles), code (AST-based chunking)
When to use it
Developers looking to add persistent memory and user context to AI products
Users who want a personal AI assistant to remember information across conversations
When you need an integrated memory solution without building RAG, vector DB, and embedding pipelines from scratch