Woolly
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
Woolly
One memory for all your AI tools. Local, private, and always up to date.
Ollama runs your models. Woolly remembers for them.
Website · Docs · Download for Mac and Windows · Bluesky · LinkedIn

Every AI app forgets you, or keeps its own memory on someone else's servers. Claude doesn't know what you told ChatGPT. Cursor doesn't know what you decided in Claude Code. And your teammates' agents start from zero every time.
Woolly gives all of them one memory. It runs on your computer, it's stored as Markdown files you can read, and nothing leaves your machine.
Get started
curl -fsSL https://usewoolly.com/install.sh | sh
woolly connectThat's it. Claude Code, Claude Desktop, Cursor, VS Code, Windsurf, Codex, Gemini CLI, LM Studio and opencode now share one memory.
Prefer an app? Download Woolly for Mac or Windows. Also on Homebrew and Linux.
Related MCP server: flux7-memory
Your team's AI shares one project memory
woolly initRun it in a repo and commit .woolly/. Now every teammate's agent reads and updates the same project memory: the stack, conventions, decisions, gotchas and commands. It's a CLAUDE.md that keeps itself up to date.
Decisions are saved for you. When a Claude Code session ends, Woolly saves what you decided, why, and what you turned down, so no agent suggests it again.
Facts stay current. A newer fact replaces the old one instead of piling up.
Personal stays personal. Facts about you never go in the repo.
Plain Markdown in git. One fact per line, readable diffs, edit it by hand.
Why Woolly
Every app, one memory. Tell Claude you moved to Denver, and Cursor knows it too. How it works →
Memory that loads by itself. Claude Code starts every session with your profile, the project's notes and where you left off. Connect →
Always up to date. When a fact changes, the new one replaces the old. When an AI guesses something that contradicts you, your version wins and the guess is flagged for review.
Gives Ollama a memory. woolly proxy adds memory to every chat with Ollama, or any app that uses the OpenAI or Anthropic API. Proxy →
Brings your memory with you. Drop in your ChatGPT or Claude export and keep the facts that matter. Import →
Private by design. No cloud, no account, no tracking. Passwords and API keys are refused, and an AI can't plant instructions in your memory. Security →
See everything. The Woolly app shows what your AI tools know and which app saved each memory. Search, edit or forget anything. The app →
Syncs your devices. Laptop, desktop and Raspberry Pi stay in step over your own network, encrypted. Sync →
Build it into your app
npm install woolly-ai # or: pip install woolly-aiimport { Woolly } from 'woolly-ai';
const w = new Woolly({ app: 'My app' });
const system = `You are a helpful assistant.\n\n${await w.context(userMessage)}`;One line adds memory to the Vercel AI SDK too. JavaScript → · Python →
Fast and accurate, on a laptop
On a sample of the LOCOMO long-conversation benchmark, Woolly found the right memory for 87% of questions in a median of 23 ms, running entirely on a laptop with Ollama. Run it yourself →
Do I need Ollama?
No. Woolly works on its own. With Ollama running, it also searches by meaning, learns facts from your chats and saves decisions when a session ends.
Learn more
Docs and all commands
Contributing, including how the repo is laid out
MIT license.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
Maintenance
Related MCP Connectors
One memory, every AI. A shared, user-owned markdown memory your AI clients read and write over MCP.
Shared memory for your AI tools: tasks, notes and context every tool can read and write over MCP.
One memory, every AI: Claude, ChatGPT, Perplexity, Gemini, Cursor, OpenClaw, Hermes, any MCP client.
An MCP memory server. One memory your agents share — across models, devices and apps.
Related MCP Servers
- AlicenseBqualityCmaintenancePersistent memory engine for AI coding agents. Single Go binary, zero runtime dependencies, MCP-native. Stores, searches, and deduplicates memories across sessions using embedded SQLite with hybrid FTS + semantic search, memory decay, relation graph, and token-budget context assembly.1023MIT
- AlicenseNot gradedqualityBmaintenanceGoverned multi-agent memory for AI agents. Hybrid markdown + SQLite store with full-text search, vector retrieval, and LLM reranking. Three transports: MCP stdio, HTTP JSON-RPC, and MCP SSE. One Go binary1Apache 2.0
- AlicenseAqualityCmaintenanceA local-first MCP server for shared memory across AI tools, enabling context capture and resume across sessions.2663 npm6Apache 2.0
- AlicenseAqualityBmaintenanceLocal-first shared memory and task coordination for AI coding agents. One Go binary, MCP server, markdown files you own. Hooks for Claude Code and Codex CLI (and their desktop apps).306MIT