AI · Web app
supermemory
External memory and continuous learning engine for AI agents and assistants
Originally launched as an AI second-brain bookmark manager, Supermemory has evolved into an infrastructure engine providing persistent memory and continual learning for AI agents. It extracts and updates context from documents, chats, and URLs into a vector-graph database, dynamically injecting relevant memories into model prompts via API, MCP, or native integrations.
What it's for
- Provide persistent context and continuous memory across AI agent sessions
- Connect AI assistants like Claude Desktop, Cursor, and ChatGPT via Model Context Protocol (MCP)
- Index and retrieve knowledge from complex documents, raw data, and web URLs for LLMs
- Self-host an air-gapped vector-graph memory layer for enterprise AI workflows
Features
- Vector-graph database architecture that maps interconnected memories over time
- Learner-1 model that continuously extracts, merges, and updates context
- Real-time token injection into agent context windows
- Model Context Protocol (MCP) support and direct plugins for Cursor, Claude, and ChatGPT
- Advanced document parsing with OCR and diagram extraction for PDFs
- Deployable via cloud API, self-hosted Docker, VPC, or local environments
- Enterprise compliance including SOC 2, HIPAA, and GDPR alignment
Pricing
Freemium. Free tier available; paid Pro, Max, Scale, and Enterprise plans offered.
Details
- Type
- Web app
- Category
- AI
- Topic
- Agent Infrastructure
- Open source
- Yes
- Platforms
- Web, API, Self-hosted, macOS, Linux, Windows
- For
- AI engineers, developers building autonomous agents, and teams needing persistent context across LLMs.
- Alternative to
- Mem0, Zep, LangChain
- Tags
- memory, agents, mcp, vector search, knowledge base, developer tools
Why it was saved
- @limitlesswithishan: Consolidates scattered bookmarks and notes from Apple Notes, Pinterest, Chrome, Twitter, and Notion into a single searchable hub with an AI writing assistant.
Compare
Related
Enterprise platform to build autonomous AI agents and automated business workflows.
A research paper formalizing architecture and context engineering for CLI coding agents.
Open-source persistent memory layer for autonomous AI agents using knowledge graphs
Real-time web search and markdown extraction MCP server for AI agents.
Open-source tutorials and project templates for LLMs, RAG, and AI agents.
Open-weights fine-tune of Llama 3.1 built for agentic function calling.
Updated