agentmemory
Persistent memory for AI coding agents such as Claude Code, Cursor, Codex and Gemini CLI, so an agent keeps a project's context across sessions instead of being re-briefed each time.
agentmemory gives AI coding agents persistent memory across sessions, so a project's context does not need to be re-explained each time. It works with Claude Code, GitHub Copilot CLI, Cursor, Gemini CLI, Codex CLI and other MCP clients, and is built on the iii engine, extending a pattern of confidence scoring, lifecycle management, knowledge graphs and hybrid search.
It installs and configures itself with a single npx command: an interactive setup wires the chosen agents, optionally selects an LLM provider, seeds configuration, and starts a local memory server together with its pinned iii engine. Without a provider, search falls back to BM25 keyword matching; a local embedding model can be enabled instead for on-device semantic recall.
The local server serves REST and MCP over HTTP, plus a viewer, over several local ports. It is written in TypeScript and licensed under the Apache-2.0 license.