agentmemory
#1 Persistent memory for AI coding agents based on real-world benchmarks
agentmemory is a persistent memory system for AI coding agents that operates as an MCP server, providing 54 memory-related tools to agents via the Model Context Protocol. It automatically captures every tool use, session, and observation from coding agents, compresses them into searchable memory using hybrid BM25 + vector + knowledge graph retrieval, and injects relevant context at the start of each new session. The system eliminates repetitive explanations across sessions by remembering project decisions, architecture patterns, bug fixes, and preferences. It works with any MCP-compatible client including Claude Code, Cursor, Cline, GitHub Copilot CLI, Gemini CLI, and 20+ other agents through a single shared memory server.
Key Features
Use Cases
- 01Eliminating re-explanation of project architecture, authentication patterns, and tech stack choices across coding sessions
- 02Remembering bug fixes, refactoring decisions, and why specific libraries were chosen over alternatives
- 03Building searchable knowledge bases from coding sessions without manual documentation
- 04Coordinating context between multiple AI agents (architect, developer, reviewer) working on the same codebase
- 05Importing and replaying historical Claude Code JSONL transcripts to mine lessons and patterns
- 06Running keyless with BM25-only search for privacy-sensitive environments without embedding API calls
Related MCP Servers
View moremarkitdown
Python tool for converting files and office documents to Markdown.
firecrawl
The context API to search, scrape, and interact with the web at scale. 🔥
prompts.chat
f.k.a. Awesome ChatGPT Prompts. Share, discover, and collect prompts from the community. Free and open source — self-host for your organization with complete privacy.
langflow
Langflow is a powerful tool for building and deploying AI-powered agents and workflows.
agentmemory — FAQ
What is agentmemory?+
agentmemory is an MCP server that gives AI coding agents persistent, searchable memory across sessions. It automatically captures observations from tool use, compresses them into structured memory, and injects relevant context when new sessions start.
How do I install agentmemory?+
Run `npx -y @agentmemory/agentmemory@latest` to start the memory server, then add the MCP server config to your agent (Claude Code, Cursor, Cline, etc.) using the standard mcpServers block with command 'npx -y @agentmemory/mcp'. For Claude Code, install the plugin via `/plugin marketplace add rohitg00/agentmemory` and `/plugin install agentmemory`.
Which AI clients does agentmemory work with?+
agentmemory works with any MCP-compatible client including Claude Code, Cursor, Claude Desktop, Cline, Roo Code, GitHub Copilot CLI, Gemini CLI, Codex CLI, OpenCode, Goose, Kilo Code, Warp, and 20+ others. It also provides REST API access for clients like Aider.
Do I need API keys or a paid service?+
No API keys are required for basic operation (keyless mode). The system runs locally with BM25 search and optional local embeddings. LLM providers (Anthropic, OpenAI, Gemini, Ollama) are optional and only needed for LLM-written compression (when AGENTMEMORY_AUTO_COMPRESS=true) or semantic vector search.
Is agentmemory free?+
Yes, agentmemory is open source (Apache-2.0 license) and free to use. It runs locally with no external dependencies. Optional embedding and LLM providers may have their own costs.
What are the prerequisites?+
Node.js 20 or newer with npm and npx. macOS/Linux auto-install requires curl, a POSIX shell, and tar. Windows requires manual iii-engine v0.11.2 installation, WSL2, or Docker Desktop.
How do I install agentmemory?+
Open the source repository on GitHub and follow its README. agentmemory is a mcp server — MCP Agents Market links you directly to the official repo.
Is agentmemory free?+
agentmemory is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.