Every morning, re-explaining the same architectural decisions to GitHub Copilot or whatever AI coding tool your team uses wastes real time. A computer science student at the University of Toronto built a VS Code extension to stop that.
ContextSync saves your AI chat history as structured Markdown files in a folder you already sync - OneDrive, iCloud, or an Obsidian vault. When a teammate opens VS Code the next session, the context from the previous one is already there. No backend to set up, no new service to pay for.
The problem it targets is genuine. AI coding assistants like Copilot, Cursor, and Cody operate session-by-session - they have no memory of what your team decided last Tuesday, which database schema you landed on, or why you deliberately avoided a particular library. Each new chat starts blank. For solo developers that's annoying; for teams it compounds.
The approach is file-based rather than server-based: your context lives in plain Markdown in whatever sync folder you already trust. Obsidian users get an extra benefit - the files drop straight into a vault, making them searchable alongside existing notes.
This is a student side project, so temper expectations. The extension is free on the VS Code Marketplace and the codebase is early. How well it handles large histories, merge conflicts between teammates, or edge cases in different AI tools hasn't been publicly stress-tested.
That said, the core idea - using a shared file sync layer as lightweight shared memory for AI coding tools - is practical and obvious in retrospect. The fact that Copilot, Cursor, and Cody haven't built something like this natively is a genuine gap, and a file-based workaround is a reasonable stopgap while waiting for them to catch up.