I've been working on Context Sync (https://github.com/Intina47/context-sync) - an MCP server that gives AI coding assistants persistent memory across platforms and sessions.
Today I'm excited to share v0.6.0, which adds VS Code/GitHub Copilot support and major performance optimizations.
## The Problem
We've all been there: you're deep into a coding session with Claude, hit the context limit, start a new chat, and suddenly Claude has forgotten your entire project architecture. You spend 30 minutes re-explaining the same context. Every. Single. Day.
## The Solution
Context Sync gives AI assistants persistent memory. Start a project Monday in Cursor, continue Tuesday in Claude Desktop, debug Wednesday in VS Code - the AI remembers everything.
## What's New in v0.6.0
• *VS Code + GitHub Copilot support* - Now works across all major AI coding platforms
• *2-5x faster database operations* - Prepared statement caching
• *File size safety limits* - Prevents OOM crashes with 5MB max file size
• *Real-time cache invalidation* - File watchers automatically update caches
• *Async file operations* - Non-blocking I/O for better performance
## Technical Details
Built as an MCP (Model Context Protocol) server in TypeScript with SQLite. Local-first, no cloud sync required. Uses intelligent caching, file watching with chokidar, and async/await throughout.
The performance optimizations were crucial - we were seeing memory issues with large codebases, so v0.6.0 adds file size limits, async operations, and much smarter caching.
I've been working on Context Sync (https://github.com/Intina47/context-sync) - an MCP server that gives AI coding assistants persistent memory across platforms and sessions.
Today I'm excited to share v0.6.0, which adds VS Code/GitHub Copilot support and major performance optimizations.
## The Problem We've all been there: you're deep into a coding session with Claude, hit the context limit, start a new chat, and suddenly Claude has forgotten your entire project architecture. You spend 30 minutes re-explaining the same context. Every. Single. Day.
## The Solution Context Sync gives AI assistants persistent memory. Start a project Monday in Cursor, continue Tuesday in Claude Desktop, debug Wednesday in VS Code - the AI remembers everything.
## What's New in v0.6.0 • *VS Code + GitHub Copilot support* - Now works across all major AI coding platforms • *2-5x faster database operations* - Prepared statement caching • *File size safety limits* - Prevents OOM crashes with 5MB max file size • *Real-time cache invalidation* - File watchers automatically update caches • *Async file operations* - Non-blocking I/O for better performance
## Technical Details Built as an MCP (Model Context Protocol) server in TypeScript with SQLite. Local-first, no cloud sync required. Uses intelligent caching, file watching with chokidar, and async/await throughout.
The performance optimizations were crucial - we were seeing memory issues with large codebases, so v0.6.0 adds file size limits, async operations, and much smarter caching.
## Try It ```bash
npm install -g @context-sync/server