Skip to main content
MCP server CX Codex

Woolly — Claude Code MCP server: what it does, how to install

Woolly — an MCP server for Claude Code unites AI tool memory into a single private database on your laptop. Project context and decisions sync across tools in local Markdown files.

MIT ★ 13 updated 2026-11-15

Open on GitHub

What Woolly does

Woolly connects AI coding tools like Claude Code, Cursor and Codex by sharing a single folder of plain Markdown files on your laptop. Every coding session starts with your profile and project notes already loaded, while decisions and facts from the conversation are saved automatically. The server runs entirely offline without cloud storage, user accounts or telemetry, and refuses to store API keys or passwords.

The tool initializes project memory inside a repository using .woolly/ for team collaboration on stack details and conventions. Local models and API clients connect through a built-in proxy, and encrypted network sync keeps multiple machines aligned. Private imports from ChatGPT and Claude histories bring existing notes into the shared storage.

What it helps with

  • share memory across multiple AI coding assistants
  • persist stack decisions and context between sessions
  • maintain team project notes in local Markdown files
  • sync memory securely across personal machines

Example prompts for the agent

  • Connect the shared memory with woolly connect
  • Initialize repository memory with woolly init
  • Start the local memory proxy with woolly proxy

How to install Woolly

Install

The MCP registry has no launch command for it: see the repository README

Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more

Questions about Woolly

How to install Woolly in Claude Code?

Run the installation script in your terminal using curl -fsSL https://usewoolly.com/install.sh | sh followed by woolly connect. On Windows, run irm https://usewoolly.com/install.ps1 | iex instead.

What does Woolly do?

Woolly acts as a shared memory server for AI coding tools on your machine. It stores project context, choices and conventions in readable Markdown files so different assistants share the same background.

What is needed to run Woolly?

Woolly runs locally on macOS or Windows without cloud dependencies. Optional features like semantic search and chat learning work when running Ollama alongside it.

Similar MCP servers

All MCP servers for Agent setup: memory, context, orchestration →

Project sponsors

Companies that keep this analytics open