Context Sync

Finally, AI that remembers.

🎉 NEW in v1.0.3: Your AI Can Now Write Directly to Notion

Ask AI to generate feature docs → AI writes them to Notion. Pull project specs from Notion → AI understands your plan instantly.

💡 "Create API documentation in Notion" • 📄 "Export this architecture decision to Notion" • 🔍 "Read the product spec from Notion and implement it"

See What's Possible → | Full Changelog

Context Sync is the open-source infrastructure that gives AI systems persistent memory across all your development tools, sessions, and projects.

Local-first • Open source • Platform agnostic • Built by developers, for developers

Supports:

The Problem Every Developer Faces

AI systems lose context between conversations. You spend hours explaining your codebase, architecture decisions, and project context... then close the chat and start over tomorrow.

The memory loss problem affects everyone:

• Freelancers juggling multiple client projects

• Teams building complex applications

• Open source maintainers across different tools

• Anyone using AI for serious development work

This isn't just inconvenient - it's fundamentally broken.

Context Sync: GitHub for AI Memory

Context Sync creates persistent, queryable memory for AI systems across all development environments.

Think of it as distributed version control for AI context:

Every AI tool gets the same shared memory.

Why Open Source?

AI memory shouldn't be controlled by corporations.

Context Sync is built by developers who were tired of:

• Losing context between AI conversations

• Being locked into single AI platforms

• Having no control over their development data

• Paying subscription fees for basic memory functionality

Our philosophy:

• Local-first: Your data stays on your machine

• Platform agnostic: Works with any AI that supports MCP

• Community-driven: Built for developers, by developers

• Extensible: Open architecture for unlimited customization

Quick Start

That's it. Context Sync auto-configures and guides you through first-time setup with natural language instructions.

🎯 See the Problem in Action

Before Context Sync:

After Context Sync:

That's the power of persistent AI memory.

🚀 Get Started in 2 Minutes

Context Sync automatically detects your platform and guides you through setup with personalized instructions.

📖 Need detailed documentation? See the Context Sync v1.0.0 Comprehensive Documentation for complete setup guides, troubleshooting, and advanced configuration.

🔧 Platform-Specific Setup

After installing Context Sync globally:

1. Restart VS Code completely

2. Open Copilot Chat (Ctrl+Shift+I / Cmd+Shift+I)

3. Switch to Agent mode (if available)

4. Look for context-sync in the Tools list

5. Test: Ask Copilot "help me get started with context-sync"

Context Sync should appear in available tools! ✨

🔧 Setup for Cursor

After installing Context Sync, in Claude Desktop type:

Claude will give you OS-specific instructions! ✨

1. Open Cursor: Settings → MCP

2. Add this configuration:

3. Refresh MCP servers

4. Test: "Help me get started with context-sync"

Done! ✅

After installing Context Sync globally:

1. Restart Claude Desktop completely

2. Start a new conversation

3. Test: Ask Claude "help me get started with context-sync"

Claude will automatically configure itself! ✨

Windows:

macOS:

Linux:

Add this configuration:

Restart Claude Desktop and test: "Help me get started with context-sync"

After installing Context Sync globally:

1. Install Continue.dev extension in VS Code

2. Restart VS Code completely

3. Open Continue chat (Ctrl+Shift+M / Cmd+Shift+M)

4. Test: Ask Continue "help me get started with context-sync"

Context Sync should be available as an MCP tool! ✨

1. Open Continue config: ~/.continue/config.yaml

2. Add MCP server configuration:

3. Restart VS Code

4. Test: "Help me get started with context-sync"

After installing Context Sync globally:

1. Open Windsurf Settings → MCP

2. Add new server with name context-sync

3. Command: npx

4. Args: -y @context-sync/server

5. Save and restart Windsurf

6. Test: Ask Windsurf "help me get started with context-sync"

After installing Context Sync globally:

1. Open Zed Settings (Cmd+, / Ctrl+,)

2. Navigate to MCP section

3. Add server:

   • Name: context-sync

   • Command: npx

   • Args: ["-y", "@context-sync/server"]

4. Restart Zed

5. Test: Ask Zed assistant "help me get started with context-sync"

After installing Context Sync globally:

1. Open TabNine settings

2. Navigate to MCP Servers

3. Add new server:

4. Restart your editor

5. Test: Ask TabNine "help me get started with context-sync"

🔗 Notion Integration

New in v1.0.3: Sync your AI context directly to Notion!

Context Sync now includes native Notion API integration with interactive setup wizard:

Features:

• 🔍 Search your Notion workspace

• 📖 Read and update pages with proper formatting

• 📝 Create new documentation automatically

• 🎯 Export decisions as Architecture Decision Records (ADRs)

• 📊 Generate project dashboards with tech stack and architecture

• ✨ Markdown support - Beautiful, formatted pages in Notion

Benefits:

• ✅ Smart updates - Skip wizard if already configured

• ✅ User-friendly setup - Select pages by number, not UUID

• ✅ Input validation - Catches errors before API calls

• ✅ Native integration - Works seamlessly with Context Sync tools

📚 Full Notion Integration Guide →

How It Works: Distributed AI Memory

Context Sync creates a persistent knowledge layer between you and AI systems using the Model Context Protocol (MCP).

Architecture:

Key advantages:

• Universal compatibility: Works with any MCP-enabled AI

• Local ownership: All data stays on your machine

• Zero vendor lock-in: Open source and extensible

• Intelligent caching: Only loads what AI needs when it needs it

What Makes Context Sync Different

🌍 Universal AI Memory Layer

Context Sync isn't just another tool - it's infrastructure.

Like Git for code, Context Sync is version control for AI knowledge:

• Distributed: Every AI tool gets the same shared memory

• Local-first: Your data, your control, your privacy

• Platform agnostic: Works with any MCP-compatible AI

• Extensible: Open source architecture for unlimited customization

Current ecosystem support:

• ✅ Claude Desktop (Mac/Windows/Linux)

• ✅ Cursor IDE

• ✅ VS Code + GitHub Copilot

• ✅ Continue.dev

• ✅ Windsurf

• ✅ Zed Editor

• ✅ TabNine

• ✅ Any MCP-compatible AI tool

• 🔄 More platforms added regularly via community contributions

🧠 Intelligent Context Management

Smart memory that scales with your projects:

Project-Aware Context:

• Automatic project detection and initialization

• Tech stack recognition (TypeScript, React, Python, etc.)

• Architecture decision tracking with reasoning

• Code structure understanding and analysis

Efficient Memory Usage:

• 1-3K tokens per project (not full conversation dumps)

• On-demand querying (AI requests details as needed)

• Structured summaries instead of raw chat logs

• Never saturates context windows

Developer-Focused Features:

• File operations (read, write, search) with approval workflows

• Git integration (status, diffs, branches)

• Dependency analysis and call graph tracing

• TODO management with cross-project tracking

• Code symbol search and type analysis

🔒 Privacy-First Architecture

Your code stays yours:

• 100% local storage (SQLite database on your machine)

• No cloud dependencies (optional git-based sync available)

• No tracking or analytics (we don't even have servers!)

• Open source transparency (audit every line of code)

• Zero vendor lock-in (export your data anytime)

Context Sync vs. Existing Solutions

The fundamental difference: Context Sync is infrastructure, not a product.

Other approaches:

• Chat-based solutions: Store conversations → become bloated and slow

• Platform-specific tools: Lock you into one AI tool

• Cloud-based services: Your data on someone else's servers

• Proprietary solutions: No customization, no transparency

Context Sync approach:

• Memory infrastructure: Structured, queryable knowledge base

• Universal compatibility: Works with any MCP-enabled AI

• Local-first: Your data stays on your machine

• Open source: Community-driven development and customization

Complementary, not competitive:

• Use with Claude Pro for more daily messages

• Works alongside /compact for conversation compression

• Enhances existing AI tools rather than replacing them

💬 Common Questions

Honest answer: Business incentives.

If Claude remembered everything perfectly:

• You'd have shorter conversations

• Use fewer messages

• Hit rate limits slower

Persistent memory makes AI more useful but potentially less profitable for AI companies.

That's why Context Sync is:

• Open source (no profit motive)

• Local-first (you own your data)

• Community-driven (built by developers, for developers)

No! Context Sync uses only 1-3K tokens per project.

How it works:

1. Stores structured summaries (not full conversations)

2. AI queries for details on-demand via MCP

3. Never dumps everything into a new chat

Analogy:

• ❌ Bad: Loading 10GB codebase into RAM

• ✅ Good: IDE that loads files as needed

Example:

• Your 10K line project with 50 decisions

• Context Sync summary: ~1.5K tokens

• AI queries for specific details when needed

You never saturate because you're not copying conversations - you're giving AI access to a queryable database.

100% local. 100% yours.

• ✅ SQLite database on YOUR machine

• ✅ No cloud sync (unless you configure it)

• ✅ No data sent to our servers (we don't have servers!)

• ✅ Open source - audit the code yourself

• ✅ Delete anytime - just remove ~/.context-sync/

Database location:

• Mac/Linux: ~/.context-sync/data.db

• Windows: %USERPROFILE%\.context-sync\data.db

You control everything.

Yes! Available since v0.6.0! 🎉

VS Code with GitHub Copilot is now fully supported through MCP integration.

Setup instructions:

1. Install Context Sync: npm install -g @context-sync/server

2. Restart VS Code

3. Open Copilot Chat

4. Switch to Agent mode

5. Look for context-sync in Tools list

Currently works with:

• ✅ VS Code + GitHub Copilot (new!)

• ✅ Cursor IDE (full support)

• ✅ Claude Desktop (full support)

Also supported since v0.6.0!

Claude Code just launched and supports MCP, so integration should be straightforward.

We're prioritizing:

1. VS Code extension

2. Claude Code CLI

3. Better onboarding

Want it sooner? Let us know in GitHub Discussions!

Not yet. Mobile requires:

• Claude mobile app to support MCP (not available yet)

• OR a custom mobile app (planned for future)

Current workaround:

• Use Claude.ai web on mobile (read-only)

• Full features on desktop only

Mobile support depends on Anthropic adding MCP to their mobile app.

Context Sync is 100% free and open source (MIT License).

Why free?

• Built by developers, for developers

• Solves a problem we personally had

• Community-driven development

• No profit motive = no business incentives to limit features

You might pay for:

• Claude Pro subscription (recommended but not required)

• Your time (2 minutes to set up)

That's it!

Context Sync handles multiple projects beautifully!

Each project maintains its own:

• Architecture decisions

• Tech stack

• TODOs

• Code context

• Conversation history

🎬 Real-World Example

Freelance Developer Workflow

Monday Morning (Cursor):

Tuesday Afternoon (Claude Desktop):

Wednesday (Cursor):

No re-explaining. No context loss. Just continuous progress across tools.

🗺️ Roadmap

See detailed roadmap: ROADMAP.md

📊 Stats

Recent milestones:

• 🎉 26 stars in first 24 hours

• 📦 400+ npm downloads in launch week

• 🚀 #17 on Product Hunt

• 💬 20K+ Reddit impressions

🛠️ Advanced Features

Project Management

• Initialize and switch between projects

• Track architecture and tech stack

• Store design decisions with reasoning

• Manage TODOs and priorities

Code Analysis

• Dependency graph analysis

• Function call traces

• Type definition lookup

• Find all references

• Impact analysis

File Operations

• Read project structure

• Search files and content

• Modify files (with approval)

• Preview changes before applying

Git Integration

• Status and diff viewing

• Branch information

• Commit message suggestions

• Change tracking

Cross-Platform

• Seamless Cursor ↔ Claude sync

• Platform-specific optimizations

• Context handoff automation

Developer Tools

• 50+ MCP tools

• Extensible architecture

• TypeScript + SQLite

• Open source

Join the AI Infrastructure Revolution

Context Sync is built by the developer community, for the developer community.

🚀 Why Contribute?

You're not just contributing to a tool - you're building the future of AI-assisted development.

• Shape AI tooling standards: Help define how AI systems should handle persistent memory

• Solve your own problems: Build features you need for your workflow

• Learn cutting-edge tech: Work with MCP, TypeScript, SQLite, and AI integrations

• Join a movement: Be part of making AI development tools truly open and extensible

🛠️ Ways to Contribute

Code Contributions:

• Add support for new AI platforms (Gemini, Ollama, etc.)

• Implement new analysis tools (Python dependency tracking, etc.)

• Build integrations for more editors (Vim, Emacs, etc.)

• Optimize performance and memory usage

Non-Code Contributions:

• Write documentation and tutorials

• Test beta features and report bugs

• Share use cases and workflows

• Help other developers in discussions

• Create example projects and templates

Community Building:

• Share Context Sync with fellow developers

• Write blog posts about your experience

• Speak at conferences or meetups

• Contribute to roadmap planning

📋 Current Priorities

Help wanted on:

• Python ecosystem support - pip, poetry, requirements analysis

• Mobile/web integration - React Native, Expo, web development workflows

• Docker/containerization - workspace detection for containerized apps

• More Git integrations - PR analysis, commit message generation

• Performance optimizations - faster workspace scanning, better caching

• UI/dashboard development - web interface for memory management

🎯 Get Started Contributing

1. Star the repo to show support

2. Join discussions to share ideas and get help

3. Read for technical guidelines

4. Pick an issue labeled good-first-issue or help-wanted

5. Submit a PR - we review quickly and provide feedback

New to open source? We're beginner-friendly! Many contributors started here.

💬 Community Links

• GitHub Discussions: Feature requests, questions, showcase

• Issues: Bug reports and feature planning

• Pull Requests: Code contributions welcome

• Roadmap: ROADMAP.md - see what's coming next

🔧 Troubleshooting

1. Verify installation:

2. Check config file:

3. Restart Claude completely:

• Mac: ⌘ + Q (force quit)

• Windows: Right-click tray icon → Exit

4. Check MCP servers in Claude: Look for Context Sync in settings

Still stuck? Create an issue

1. Open Cursor settings: Settings → MCP

2. Verify configuration exists:

3. Refresh MCP servers in Cursor

4. Test: Ask AI "What's my current project?"

Set up a workspace first:

Or check existing projects:

1. Check platform status:

2. Verify both platforms are configured

3. Try manual sync:

More help: TROUBLESHOOTING.md

License & Philosophy

MIT License - Use commercially, modify freely, redistribute openly.

Context Sync is truly open source:

• No dual licensing schemes

• No "enterprise" vs "community" versions

• No feature paywalls or subscription tiers

• No proprietary extensions or locked ecosystem

Why MIT? Because AI tooling infrastructure should belong to the developer community, not corporations.

The Vision

We're building the git of AI development.

Just as git transformed how developers collaborate on code, Context Sync is transforming how developers collaborate with AI systems.

Current state: AI tools are isolated, forgetful, and platform-locked Our vision: Universal, persistent, extensible AI memory layer End goal: AI that truly understands and remembers your development context across all tools and platforms

This is bigger than Context Sync - we're establishing standards and protocols that any tool can implement. Think of us as the Apache Foundation for AI development infrastructure.

Star History & Community

Growing fast thanks to developers like you:

Recent milestones:

• 🎉 26 stars in first 24 hours

• 📦 400+ npm downloads in launch week

• 💬 Active community in GitHub Discussions

• 🚀 Growing contributor base from 6 countries

Spread the Word

Help other developers discover Context Sync:

• ⭐ Star the repository (most important!)

• 🐦 Share on social media with your experience

• � Write about it in blogs, newsletters, forums

• 💬 Tell colleagues about the productivity gains

• 🎤 Present at meetups or conferences

• 🤝 Contribute code or documentation

Every developer who discovers Context Sync makes the AI development ecosystem a little more open and powerful.

Built by developers, for developers. Join the movement.

⭐ Star on GitHub • 📖 Documentation • 💬 Discussions • 🐛 Issues