io.github.Intina47/context-sync

Official

by Intina47 · TypeScript

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

Syncs AI context across Claude, VSCode, Cursor, Continue, Windsurf, and Zed.

Topics

developer-toolsmcp-serversqliteworkspacecontext-synccontext-managementmodel-context-protocolproductivitymemoryaiai-memoryclawdbotmoltbotopenclaw

README.md3 KBView on GitHub

Context Sync

Local-first project memory for AI coding tools over MCP.

Context Sync gives an agent a persistent memory layer for a codebase:

project identity and tech stack
active work, decisions, constraints, and caveats
structured file exploration
git-aware context
optional read-only Notion lookup

Context Sync

What It Does

Context Sync is built for session continuity. Instead of relying on a model to remember earlier conversations, it stores the project context that matters and makes it retrievable through a small MCP tool surface.

The current core tools are:

set_project
remember
recall
read_file
search
structure
git
notion

Install

Install globally:

bash

npm install -g @context-sync/server

Auto-configuration runs during global install. After install, restart your AI tool.

If you install locally instead of globally, auto-config does not run. Use the manual config guide in docs/CONFIG.md.

Verify

Check that context-sync appears in your MCP tool list:

Claude Desktop: open the tools list
Cursor: open the tools list
VS Code / Copilot Chat Agent mode: check tools
Continue.dev: open the Continue panel
Codex CLI: codex mcp list
Claude Code: claude mcp list

If auto-config fails, use docs/TROUBLESHOOTING.md and docs/CONFIG.md.

Optional Notion Setup

Run the setup wizard if you want Notion integration:

bash

context-sync-setup

or:

bash

npx context-sync-setup

Notion support is intentionally read-only at the tool surface:

notion({ action: "search", query: "..." })
notion({ action: "read", pageId: "..." })

First-Time Workflow

Start every new session by setting the project first:

text

1. set_project({ path: "/absolute/path/to/project" })
2. recall()
3. structure({ depth: 2 })
4. read_file({ path: "src/index.ts" })
5. remember({ type: "decision", content: "Use SQLite for local storage" })

Recommended pattern:

set_project once per project/session start
recall to recover context
structure and search to explore
read_file for focused inspection
remember when something should survive the current session

Git Integration

When you run set_project inside a git repository, Context Sync can install hooks for automatic context capture.

Installed hooks:

post-commit
pre-push
post-merge
post-checkout

Hooks are marked with Context Sync Auto-Hook, and existing hooks are backed up first.

The git tool supports:

status
context
hotspots
coupling
blame
analysis

Data And Config

Default locations:

Database: ~/.context-sync/data.db
Config: ~/.context-sync/config.json
Install status: ~/.context-sync/install-status.json

Custom database options:

CLI: context-sync --db-path /absolute/path/to/db
Env var: CONTEXT_SYNC_DB_PATH

See docs/DATA.md for storage details.

Supported Platforms

Claude Desktop
Cursor
VS Code + GitHub Copilot
Continue.dev
Zed
Windsurf
Codeium
TabNine
Codex CLI
Claude Code
Antigravity

Docs

Install

claude_desktop_config.json

{
  "mcpServers": {
    "context-sync": {
      "command": "npx",
      "args": [
        "-y",
        "@context-sync/server@1.0.1"
      ]
    }
  }
}