Hey Product Hunt 👋 I’ve been using Agent Replay locally for a while now before deciding to share it. The problem: agents now write most of my code, and reviewing it in a classic diff view is exhausting. Too many files, no order, no context for why a change was made. Agent Replay reads the session logs your agent already writes and plays them back like a recording. You see each change made in order, with the agent’s explanation, and an Evidence tab that shows which checks actually ran and passed (and flags deleted tests or failures hidden behind | tail). It also helps when you’ve built a side project mostly with AI and need to finally understand it: replay the git history, or have your agent turn the repo into a step-by-step course. Everything runs locally. MIT licensed. Try it with: npx agent-replay-studio@latest I’d love to hear how you review agent code today and what’s missing.

Agent-replay

A better way to review AI code

A better way to review AI code