Supermemory

// Overview

Supermemory is a memory and context layer for AI applications. AI assistants normally forget everything between conversations; Supermemory stores that context for you. It automatically extracts facts from conversations, builds and maintains user profiles, handles knowledge updates and contradictions, forgets expired information, and returns the right context when an agent needs it.

It is aimed at developers building AI products who want memory, retrieval-augmented generation (RAG), and connectors without wiring up a vector database, embedding pipeline, or chunking strategy themselves. You add memories and search them through a single API, available as an npm package for TypeScript and a PyPI package for Python. There is also an MCP server so assistants like Claude Code, Cursor, and Windsurf can read and write memory directly.

As a memory framework, Supermemory bundles the whole context stack into one system: fact extraction, hybrid search that combines RAG documents with personalized context, multi-modal file extractors, and connectors for sources such as Google Drive, Gmail, Notion, and GitHub. You can use the hosted API or self-host it locally, including fully offline with a local model through Ollama.

// What it does

• Automatic memory: extracts facts from conversations and handles temporal changes, contradictions, and automatic forgetting
• Auto-maintained user profiles that combine stable facts with recent activity, returned in a single call
• Hybrid search that pulls RAG documents and personalized memory together in one query
• Connectors for Google Drive, Gmail, Notion, OneDrive, and GitHub with real-time webhook sync
• Multi-modal extractors for PDFs, images (OCR), videos (transcription), and code (AST-aware chunking)
• Self-hostable as a single binary, with the option to run fully offline using Ollama

// Getting started

You can call the hosted API from TypeScript or Python, wire it into an AI client through MCP, or run it yourself locally. The quickest path for app builders is the SDK.

npm install supermemory
export SUPERMEMORY_API_KEY="YOUR_API_KEY"

import Supermemory from "supermemory";

const client = new Supermemory();
const USER_ID = "dhravya";

// Add memory
await client.add({
  content: "User conversation content here",
  containerTag: USER_ID,
});

// Search memories
const profile = await client.profile({
  containerTag: USER_ID,
  q: "search query",
});

npx -y install-mcp@latest https://mcp.supermemory.ai/mcp --client claude --oauth=yes

curl -fsSL https://supermemory.ai/install | bash

Commands and code are distilled from the project's own documentation — always check the official repo for the latest.

// When to use it

• Give an AI agent persistent memory of a user's preferences, projects, and past conversations across sessions
• Add RAG over your own documents and personalized user context to a product through one API instead of building a vector pipeline
• Sync knowledge from Google Drive, Gmail, Notion, or GitHub into a single searchable memory layer
• Give coding assistants like Claude Code, Cursor, or Windsurf long-term memory through the MCP server

// How Supermemory compares

Supermemory alongside other open-source agent memory tools AI/TLDR tracks, ranked by GitHub stars.