rohitg00/agentmemory is a persistent memory server that keeps the information a coding agent has learned available for the next task. It is written in TypeScript and follows the Apache-2.0 license. The latest release is v0.9.30, and the given records state no changes for this version.
Persistent memory works by having the agent leave what it learned on the server and pull it out again in a later task. agentmemory is built on the iii engine, and its README explains how to connect MCP (Model Context Protocol) clients such as Claude Code, Copilot CLI, Cursor, Gemini CLI, Codex CLI, Hermes, OpenClaw, pi, and OpenCode. Installation requires Node.js 20 or later and npm-family tools. The repository contains installation, quick start, benchmark, how it works, MCP, and API guides along with automatic check badges, and it also holds READMEs in several languages. The repository was created on February 25, 2026 and has 29702 GitHub stars.
That said, the repository description introduces itself as the number one coding agent memory based on benchmarks in real environments. The README does list benchmark items, but the given records do not include the items or conditions of the benchmark that decided that ranking, nor do they explain what the iii engine provides. Of the evaluation items, only the momentum score is filled in, at 8.4 out of 10, while the freshness, technology, and quality scores are empty. So what can be confirmed right now is the set of integration targets and the installation requirements; the ranking and performance can only be judged once more material has accumulated.