AgentMemory gives AI coding agents a persistent memory, so what your agent learned in one session is still there in the next โ no more re-explaining your codebase every morning.
AgentMemory is an open-source project on GitHub with around 30,000 stars and active development in 2026. It runs as a local memory layer for coding agents such as Claude Code, Codex, Copilot, Cursor, Hermes, and OpenClaw. As you work, it captures decisions, patterns, and project context, then recalls the relevant pieces automatically in future sessions. The project positions itself on real-world benchmark results for recall quality, and because it runs locally as an MCP-compatible server, your project’s memory stays on your machine rather than in a vendor cloud.
Because the memory layer is agent-neutral, a developer can switch between Claude Code in the morning and Cursor in the afternoon and keep the same accumulated project memory. Memories are stored in structured, human-readable form, so teams can review what the agent believes about the codebase, correct mistakes, and even commit shared memory to version control alongside the code it describes.
AgentMemory is free and open source under the Apache-2.0 license. You run it locally alongside your existing agent and model subscriptions; there is no hosted tier required to use it.
Developers who live in AI coding agents and are tired of repeating the same project context โ teams with long-lived codebases benefit most, since accumulated memory compounds across every session.