LaReview turns a GitHub PR or unified diff into a structured review plan. Instead of scrolling files, you review intent, tasks, and relevant hunks in a deliberate order. Everything runs locally using your own GitHub CLI and AI agent.
Hey Product Hunt 🥒
I built LaReview because the DX of code review has barely changed in years. We still scroll diffs, jump between files, and hold too much context in our heads. Most new AI review tools just add more comments on top of the same old workflow.
I wanted to try a different approach. Instead of layering more output onto the diff, LaReview restructures the review itself. A PR or unified diff is turned into a task tree, so you start with intent and then validate changes step by step, only seeing the hunks that matter for the current task.
Another big motivation was cost and control. If I already run a coding agent locally, I do not want to pay again for a bundled AI. LaReview uses the agent you already have via ACP and fetches PRs through your own GitHub CLI. No servers, no uploads, no double billing.
LaReview is fully open source and built as a cross platform desktop app using Rust. This is an early alpha and I would love feedback, especially from people who review large or complex PRs. Happy to answer questions.
running code review through an AI agent is solid. whether it blocks or just suggests matters a lot - one is a tool, the other is actually an authority in the workflow.
