Supermemory

Supermemory provides a context cloud for AI agents, offering built-in memory, retrieval-augmented generation (RAG), user profiles, connectors, and extractors. It solves the problem of agents lacking persistent, structured memory and context across sessions. The product targets developers and teams building AI applications, as well as individual power users. It is positioned as a low-latency, self-hostable API that works with any model, differentiating itself from traditional vector databases by providing a knowledge graph that evolves and merges information.

Key features

• Memory and continual learning
• SuperRAG retrieval with hybrid search
• Filesystem mount for agents
• User profiles for context
• Connectors for Slack, Notion, Drive
• Extractors for PDFs, web pages, audio
• Qualitative analysis and clustering
• Low latency (<300ms recall)
• Self-hostable and SOC 2 Type II