I built Pikai after using Cursor and Claude Code heavily in real production systems. These tools work well within a single repository, but once you’re working across multiple repos, microservices, and shared libraries, they start making changes with incomplete assumptions about the rest of the system.

Pikai provides a way to share relevant code and architectural context across repositories so tools like Cursor and Claude Code have a more accurate view of the system they’re modifying. The goal is fewer incorrect edits when changes span services.

This is an early beta shaped by day-to-day production use. I’m looking for developers working in multi-repo or microservice environments who want to try it and give direct feedback.

Happy to answer technical questions about how it works.