MemMachine

MemoryApache-2.0

MemMachine is an open-source memory layer for AI agents and LLM applications. It stores working, episodic and profile memory so an agent can recall facts and preferences from earlier sessions, across agents and models. Developers reach it through a REST API, a Python SDK, a TypeScript SDK or an MCP server. Episodic memory sits in a Neo4j graph database and profile memory in SQL, so deployment means running the server yourself in a private cloud or on-premises.

The client package needs a running MemMachine server, and the README points to a quick start guide for Docker, self-hosted and cloud options.

Pinout

  • MCP server: supported, verified
  • HTTP API: supported, verified
  • command line: not checked yet
  • Python SDK: supported, verified
  • TypeScript SDK: supported, verified
  • OpenAPI description: not checked yet
Where each pin comes from
InterfaceStatusSource we readRead on
MCP serverSupported, verifiedGitHub README, MemMachine/MemMachine, Install
HTTP APISupported, verifiedGitHub README, MemMachine/MemMachine, Interfaces
Command lineNot checked yet——
Python SDKSupported, verifiedGitHub README, MemMachine/MemMachine, Install
TypeScript SDKSupported, verifiedGitHub README, MemMachine/MemMachine, Interfaces
OpenAPI descriptionNot checked yet——

Facts

Facts about MemMachine
PurposeAn open-source memory layer for AI agents and LLM apps. GitHub README, MemMachine/MemMachine, read .
MCP commandsmemmachine-mcp-stdio for stdio mode and memmachine-mcp-http for HTTP mode. GitHub README, MemMachine/MemMachine, Install, read .
InterfacesPython SDK, TypeScript SDK, RESTful API and MCP server. GitHub README, MemMachine/MemMachine, Interfaces, read .
StorageNeo4j graph database for episodic memory and SQL for profile memory. GitHub README, MemMachine/MemMachine, Database dependencies, read .
Memory typesWorking, episodic and profile memory. GitHub README, MemMachine/MemMachine, read .
DeploymentThe open-source version can be deployed in a private cloud or on-premises. memmachine.ai home page, FAQ, read .
Enterprise versionThe site says an Enterprise version is coming soon. memmachine.ai home page, FAQ, read .
LicenseApache-2.0. Repository LICENSE file, read .
Last commit. GitHub API, latest commit on the default branch of MemMachine/MemMachine, read .

Strengths

  • Python, TypeScript, REST and MCP interfaces are documented.
  • Memory is split into working, episodic and profile types.
  • The project is Apache-2.0.

Alternatives to MemMachine

All memory tools for agents
  1. Agent memory system with retain, recall and reflect operations.

    Memory
  2. Memory layer that wraps LLM clients to store and recall agent state.

    Memory
  3. Context layer for agents, built on a temporal context graph.

    Memory
  4. Memory API with a hosted MCP server, Python and TypeScript SDKs, and a local server.

    Memory

Last check: . Method: public repository and docs read automatically, page and day written down. Next check due within 90 days. How we check

Is this your tool? Write to hello@agent-tools.directory from an address on its domain to correct a pin or a fact.