What is Agent Standup?
Agent Standup turns local session files from Claude Code, Codex, Mistral Vibe and Claude Cowork into a live call where developers can follow AI coding work. Agents appear as participants, with status, commands and file changes visible as sessions run. Past sessions can be replayed, too. It runs on your machine, is free and open source, and can be used from VS Code, compatible editors, a terminal, a desktop app or inside Claude Code.
Agent Standup Features:
Live Agent Call: The app presents active agent sessions as a meeting, with sub-agents joining when assigned work and leaving when they finish.
Agent Status: Each agent is marked as working, waiting on someone, needing your attention or done, and the view shows who is waiting on whom.
Commands and Diffs: See the command an agent runs alongside its actual output, view file changes as diffs and identify failed commands.
Session Replay: Scrub through past sessions and jump to the first error, longest wait or a looping section.
Session Recap: At the end of a replay, review a recap of agents, commands, errors, token use and changed files.
Voices and Characters: Agents can speak with natural voices or an Animal Crossing-style babble, with lip-sync, and you can add characters from images in a folder.
Multiple Ways to Use: Open the call from VS Code, Cursor, Windsurf, VSCodium, a browser launched from the terminal, a desktop app or a Claude Code plugin.
Agent Standup Benefits:
Less Log Checking: A shared view of active work reduces the need to repeatedly open separate session logs to find out what each agent is doing.
Timely Intervention: Clear waiting and attention states make it easier to notice when an agent is blocked or needs a decision from you.
Quicker Debugging: Seeing command output, diffs and marked failures together can help you locate the point where a task went wrong.
Easier Review: Replaying a session and scanning its recap gives you a practical way to understand completed work without reading every log line.
Local Operation: Because the app runs on your machine and uses no model of its own, following sessions does not consume its own tokens or credits.
Use Cases:
Monitor Parallel Tasks: Keep track of several coding agents working on separate tasks, and see when one is waiting for another to finish.
Review a Failed Run: Replay a coding session after an error, jump to the failed command and inspect the output and related file changes.
Check Completed Work: Use the session recap to review the commands an agent ran, any errors and the files it changed before moving on.
Watch Agent Briefings: Follow a lead agent as it assigns work to sub-agents and see their contributions appear in the call as they work.
Work Across Editors: Open the same local session view from VS Code, a compatible editor, the terminal or the desktop app, depending on your workflow.