Devtoolwatch Boardagentmemory review
agentmemory is a local memory server that gives AI coding agents persistent memory across sessions through hooks, MCP, or a REST API.
TL;DR
agentmemory captures what a coding agent does, compresses it into searchable memory, and injects relevant context when the next session starts. It is for developers using Claude Code, Cursor, Codex CLI, Gemini CLI and other MCP clients. All agents share one self-hosted memory server, and agentmemory can run keyless with BM25 search.
- 30kGitHub stars
- +219new stars, week of Oct 25
- ~32.6kdownloads a week, +174% vs 4-week avg
- #9in MCP and agent tools this week
agentmemory at a glance
- Category
- MCP and agent tools
- Maintained by
- Rohit Ghumare (individual, @rohitg00 on GitHub)
- License
- Apache 2.0
- Language
- TypeScript, HTML, JavaScript
- Repository
- github.com/rohitg00/agentmemory
- Website
- agent-memory.dev
- npm package
- @agentmemory/agentmemory 0.9.30
- Latest release
- v0.9.30 on GitHub, Oct 6, 2026
- First published
- Feb 25, 2026
What agentmemory does
- agentmemory captures agent activity through hooks, such as 12 hooks for Claude Code and 7 for Cursor.
- agentmemory searches memory with BM25, vector, and graph results combined through RRF fusion.
- agentmemory shares one memory server across all connected agents, so findings from one tool are available in another.
- agentmemory manages memory with 4-tier consolidation, decay, and auto-forget.
- agentmemory ships 17 skills, including /lesson, which turns corrections into confidence-weighted rules.
- agentmemory includes a real-time viewer on port 3113.
- agentmemory can isolate memory per agent with AGENTMEMORY_AGENT_SCOPE=isolated.
How it works
agentmemory runs as a local Node.js server on a pinned iii-engine v0.22.1, with SQLite storage. The first npx run picks agents to wire, an optional LLM provider, and starts the server. Agents connect through native plugins, hooks, MCP, or REST on port 3111. State is stored outside the repository in a platform data directory. The REST API requires a bearer secret by default.
How to install and start using agentmemory
With npx
npx -y @agentmemory/agentmemory@latestGlobal install with npm
npm install -g @agentmemory/agentmemory@latestAdd agent skills
npx skills add rohitg00/agentmemory -yRequirements
- Node.js 20 or newer with npm and npx.
- curl, a POSIX sh, and tar for automatic iii-engine installation on macOS and Linux.
- On native Windows, the pinned iii-engine v0.22.1 iii.exe must be installed manually; WSL2 or Docker Desktop also work.
- Hand-written REST requests to port 3111 need the bearer secret from ~/.agentmemory/secret.
First steps
- Run npx -y @agentmemory/agentmemory@latest and pick your agents and an LLM provider, or stay keyless.
- Run npx -y @agentmemory/agentmemory@latest demo to seed sample sessions and test recall.
- Set EMBEDDING_PROVIDER=local in ~/.agentmemory/.env and restart to enable on-device semantic recall.
- Use agentmemory connect to wire additional agents later.
Who agentmemory is for
- Developers who use several coding agents and want them to share one memory.
- Developers tired of re-explaining project architecture and preferences each session.
- Teams that want memory kept on their own machine with no account or cloud.
Limitations to know
- agentmemory keyless mode disables vector embeddings, so semantic queries can return zero results until an embedding provider is configured.
- agentmemory pins iii-engine v0.22.1 and will not attach to a different engine version.
- agentmemory uses four local ports: 3111, 3112, 3113, and 49134.
- agentmemory LLM-written observation compression requires both a provider and AGENTMEMORY_AUTO_COMPRESS=true.
Is agentmemory growing? Usage data
In the week of Oct 25–31, 2026, agentmemory ranked #93 on the Devtoolwatch Board and #9 in MCP and agent tools, with a momentum score of 65 out of 100. Its main package had about 32.6k downloads that week, +174% against its own 4-week average.
Show the weekly numbers
| Week of | New stars |
|---|---|
| Aug 9, 2026 | 330 |
| Aug 16, 2026 | 265 |
| Aug 23, 2026 | 495 |
| Aug 30, 2026 | 338 |
| Sep 6, 2026 | 340 |
| Sep 13, 2026 | 257 |
| Sep 20, 2026 | 304 |
| Sep 27, 2026 | 254 |
| Oct 4, 2026 | 206 |
| Oct 11, 2026 | 359 |
| Oct 18, 2026 | 260 |
| Oct 25, 2026 | 219 |
The Board ranks tools by four signals each week: new stars, new stars for the repo's size, added downloads and download growth beyond the market. How the ranking works.
agentmemory alternatives
Other MCP servers and agent tools on the Board in the week of Oct 25–31, 2026, by momentum:
| Rank | Tool | Stars | Downloads/wk |
|---|---|---|---|
| #1 | REAREA is an MCP server and CLI that lets coding agents reverse engineer native binaries, JavaScript and Electron apps, .NET assemblies,… | 415.5k | 760.6k |
| #2 | ha-mcpha-mcp is an unofficial Model Context Protocol server that lets AI assistants control and manage Home Assistant using natural language. | 5.4k | 118.1k |
| #3 | TregTreg is a registry and proxy that gives AI agents one token for a catalog of paid third-party API endpoints and team-shared tools. | 6.9k | 6.7k |
| #4 | Chrome DevTools MCPChrome DevTools MCP is an MCP server and CLI that lets coding agents control and inspect a live Chrome browser using Chrome DevTools. | 54.6k | 2.7M |
| #5 | codebase-memory-mcpHigh-performance code intelligence MCP server. Indexes codebases into a persistent knowledge graph — average repo in milliseconds. 162… | 47.9k | – |
| #6 | open-computer-use👾 Open Computer Use – Open-Source Alternative to Codex Computer Use | 2.8k | – |
Recent agentmemory releases
agentmemory FAQ
Is agentmemory free and open source?
Yes. agentmemory's source code is on GitHub (rohitg00/agentmemory) under the Apache 2.0 license, so you can use it for free.
How do I install agentmemory?
Run npx -y @agentmemory/agentmemory@latest. It needs Node.js >=20.0.0. The install section above has the other options from the project's README.
Is agentmemory still maintained?
Yes. The rohitg00/agentmemory repository was last updated on Oct 21, 2026, and the latest release, v0.9.30 on GitHub, came out on Oct 6, 2026.
How popular is agentmemory?
agentmemory has 30k GitHub stars as of Nov 1, 2026. It gained 219 new stars in the week of Oct 25–31, 2026, against a 12-week average of 302 a week. Its main package had about 32.6k downloads that week, +174% against its own 4-week average. That made it #9 among MCP servers and agent tools on the Devtoolwatch Board.
agentmemory vs Claude-Mem: how do they compare?
agentmemory and Claude-Mem are both tracked by Devtoolwatch. agentmemory has 30k GitHub stars and Claude-Mem has 107.8k. In the week of Oct 25–31, 2026, agentmemory's main package had 32.6k downloads and Claude-Mem's had 34.4k. Claude-Mem is a persistent memory plugin that captures coding agent sessions, compresses them with AI, and injects relevant context into future sessions. Claude-Mem review.
What are the best alternatives to agentmemory?
Other MCP servers and agent tools on the Devtoolwatch Board in the week of Oct 25–31, 2026, by momentum: REA (#1), ha-mcp (#2), Treg (#3), Chrome DevTools MCP (#4), codebase-memory-mcp (#5). The alternatives table above has their numbers.
Does agentmemory need an API key?
No. agentmemory runs keyless by default and uses BM25 keyword search. For free semantic recall, set EMBEDDING_PROVIDER=local, which downloads all-MiniLM-L6-v2 once and runs inference locally. An LLM provider is optional.
Which agents does agentmemory work with?
agentmemory works with any agent that supports hooks, MCP, or REST. Listed agents include Claude Code, Codex CLI, GitHub Copilot CLI, Cursor, OpenCode, Devin, Gemini CLI, Zed, Cline, Continue, Goose, and Aider via REST.
Where does agentmemory store data?
agentmemory stores state on your machine, outside the repository: ~/Library/Application Support/agentmemory on macOS, ~/.local/share/agentmemory on Linux, and %APPDATA%\agentmemory on Windows. Override it with --data-dir or AGENTMEMORY_DATA_DIR.
Does agentmemory send data to the cloud?
No, not by default. agentmemory keeps memory on your machine and is self-hosted by default. Data goes to an external service only if you configure an LLM provider.
How is agentmemory different from CLAUDE.md or .cursorrules?
agentmemory captures sessions automatically and retrieves relevant context. The README says built-in files like CLAUDE.md cap out at 200 lines, go stale, and load everything into context.
Sources
agentmemory README badge
Maintainers can show the tool's weekly Board rank in their README:
[](https://index.devtoolwatch.com/tools/rohitg00-agentmemory/)How this page is made: Devtoolwatch writes each review from the project's README, releases, package metadata and website, with an AI model, then checks every claim against the source it cites and drops what the sources don't support. The numbers come from Devtoolwatch's daily GitHub snapshots, GitHub's stargazer history, and the npm and PyPI download APIs. Spotted an error? Email [email protected].