About Remem

Remem is a self-hosted state and context database for persistent AI agents. It preserves facts, relationships, and lifecycle-aware context across sessions, so an agent can work from durable state rather than a transient conversation history.

Remem 0.5 is rewritten in Go and runs its REST API, Pebble-backed storage engine, and MCP server as one service. Choosing stable storage lets the project focus on memory, retrieval, lifecycle, and relationship features instead of maintaining another storage engine. It does not require PostgreSQL, Neo4j, Qdrant, or another external database to get started. Run it with docker compose up --build -d on infrastructure you control.

What Remem is built for

  • Persistent agent state — model facts, preferences, decisions, observations, and task context as durable data rather than prompt fragments.
  • Multiple retrieval paths — semantic, keyword, hybrid, and graph retrieval address different ways an agent may need to recover context.
  • Lifecycle controls — TTL expiration, importance decay, promotion, soft deletion, and archiving make validity and forgetting explicit.
  • Relationships and time — typed graph connections and time-windowed queries help an agent reason about context beyond the nearest semantically similar text.
  • Interfaces for agents and applications — MCP tools and a documented REST API make Remem accessible to agent runtimes and conventional services.
  • Reliable operation — Pebble storage, durable background jobs, migrations, snapshots, and repair workflows make the single-node foundation easier to run and recover.

From Community to Enterprise

The Community Edition is a complete single-node foundation: core data capabilities, tenant-aware storage, local persistence, MCP, and APIs. Because the Go implementation owns a clear storage and service boundary, future Enterprise capabilities can focus on operating Remem as organisation-wide infrastructure, including multi-tenancy, clustering, high availability, access control, audit, and a control plane.

Open source

Remem is source-available under theApache 2.0 license. You can use it freely — including in commercial products — with one restriction: you cannot offer Remem as a competing managed service. Every release converts to Apache 2.0 two years after its release date.

The source is on GitHub. Bug reports, questions, and contributions are welcome.