I built r3 to scratch my own itch when reviewing long design doc and planning doc from AI. The chat interface is inherently bad at tracking multiple pieces of feedback across different parts of a document, so I created a lightweight local web UI to streamline the review process.

I have been using it on my own projects for two weeks, and it has made design review much easier. I thought others working with AI coding agents might find it useful too.

Previously https://news.ycombinator.com/item?id=48855670. Resubmit as lots of significant improvements have been made and I consider it's near the stable state.

I'd love to hear what you think and any feedback you have!