Thinking about building a tool that records AI coding agent sessions (the prompts, edits, terminal output, decisions) into a shareable replay link — like a scrubbable timeline you could send a teammate instead of screen-recording or narrating what happened. Aimed at teams doing code review or onboarding new devs to an AI-assisted codebase.
Trying to validate before I sink time into it. Genuinely curious:
- When you review or explain an AI agent's session to someone else today, what do you actually do — screen share, paste transcripts, screenshots?
- If you've tried sharing an agent session with a teammate, what got lost or was hardest to convey?
- For onboarding specifically, would seeing how an agent solved something (not just the final diff) actually change how someone learns the codebase, or is the diff enough?
Top comments (0)