Supermemory

Supermemory is a memory layer and context engineering platform for developers, teams, and enterprises building AI agents and apps that need long-term, cross-session context.

Key capabilities include:

• Universal memory API that handles storage, chunking, retrieval, and context injection so agents can keep persistent memory across sessions.
• Five-layer context stack for production use cases, including user profiles, a memory graph, hybrid retrieval (vector + keyword) with <300ms recall, multi-modal extractors (PDFs, web pages, images, audio), and connectors (Notion, Slack, Google Drive, S3, Gmail, plus custom sources).
• Model-agnostic integration via OpenAI-compatible clients and SDKs (TypeScript, Python, REST), plus integrations with tools and frameworks such as LangChain, LangGraph, Vercel AI SDK, CrewAI, Zapier, and n8n.
• Enterprise deployment options including self-hosting, SOC 2 support, and data export controls (no training on customer data stated on the site).

Pricing: Freemium with tiers, Free ($0), Pro ($19/mo), Scale ($399/mo), and Enterprise (custom), plus overage pricing.

Notable for combining retrieval, extraction, connectors, and a graph-based memory engine in one platform, with production scale claims of 100B+ tokens processed monthly and a personal app and plugins that share one memory across tools.