# context-sync [Health: Active]

**Category:** 🧠 Knowledge & Memory  
**Repository:** https://github.com/Intina47/context-sync  
**GitHub Stars:** 190  
**npm Downloads (last month):** 384  
**Views:** 3  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/context-sync

## Description
Universal AI Memory - Sync context across Claude, VsCode, Cursor, Continue, Windsurf, Zed & more

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "context-sync": {
    "command": "npx",
    "args": ["-y","context-sync-setup"],
    "env": {
      "CONTEXT_SYNC_DB_PATH": ""
    }
  }
}
```

**Requires environment variables:** `CONTEXT_SYNC_DB_PATH` — the values above are empty placeholders; fill in real credentials before running (see the repository for what each one is for).

## Documentation

## What context-sync MCP server does

The context-sync MCP server provides a local memory layer for AI-assisted software development. It keeps project-specific information available across sessions, including the project identity, technology stack, current work, decisions, constraints, and caveats. This helps an agent recover relevant context after a new conversation or when switching between supported coding tools.

The server also exposes repository inspection capabilities. Agents can request a structured view of project files, read selected files, search project content, and query Git-related information. An optional Notion integration adds read-only search and page retrieval.

## How it works

A session normally begins with `set_project`, which establishes the project path. The agent can then call `recall` to retrieve stored context, use `structure` and `search` to inspect the repository, and call `read_file` for focused file access. Information intended to persist beyond the current session is saved with `remember`, using categories such as decisions or other project context.

When the selected project is inside a Git repository, the server can install hooks for `post-commit`, `pre-push`, `post-merge`, and `post-checkout`. Existing hooks are backed up before installation, and the added hooks are marked with `Context Sync Auto-Hook`. Git analysis includes status, repository context, hotspots, coupling, blame, and general analysis actions.

The context-sync MCP server stores its database, configuration, and installation status under `~/.context-sync` by default. A different database location can be selected through the command line or `CONTEXT_SYNC_DB_PATH`.

## Setup and configuration

Install the package globally with npm:

```bash
npm install -g @context-sync/server
```

Global installation performs automatic configuration. After installation, restart the AI coding application and verify that `context-sync` appears in its MCP tools. Local installation does not run the automatic configuration step, so it requires the manual configuration instructions referenced by the project.

Notion support is optional. Run `context-sync-setup` or `npx context-sync-setup` to configure it. The available Notion operations are limited to searching and reading pages; the provided material does not describe write operations.

## Tools and capabilities

The listed MCP tools are:

- `set_project` for selecting the active project.
- `remember` for saving durable project information.
- `recall` for retrieving stored context.
- `read_file` for focused file inspection.
- `search` for repository exploration.
- `structure` for viewing the project layout.
- `git` for repository status and analysis actions.
- `notion` for optional read-only Notion search and page reads.

Supported clients include Claude Desktop, Cursor, VS Code with GitHub Copilot, Continue.dev, Zed, Windsurf, Codeium, TabNine, Codex CLI, Claude Code, and Antigravity. The context-sync MCP server is therefore suited to workflows that move between multiple AI coding environments while keeping project context in a local store.

## Limitations and notes

Notion access is read-only at the tool layer. Automatic configuration applies to global installation, while local installation requires manual setup. The server depends on the agent setting a project before repository-specific context can be recalled or explored. The provided material does not specify a license, authentication requirements, or remote hosting option.

_Full upstream README: https://allmcps.com/mcp/context-sync/readme_