- Published on
AI agents in a stand-up call
- Authors
- Name
- DeepHorus
Agent Standup turns a Claude Code session into a live stand-up call. The lead gives orders, sub-agents join, share their screen, and report back. The tool is described as a bit silly and weirdly useful. It is free and open source.
It works wherever Claude Code runs. A small local server offers several ways in. Users can access it through Cursor, Windsurf, or VSCodium using the same extension installed from a VSIX file. Any terminal works without an editor.
The desktop app provides its own window and a menu bar indicator, though it is not signed by Apple. A Claude Code status line and tmux option give a one-line indicator. The main prerequisite is that Claude Code must have been used on the machine at least once.
The live call shows one status per agent. Sub-agents join, share their screen, and leave. It includes a context gauge and token count. Replay, recap, tree, and video features are available. Voices, custom characters, and 8 languages are supported.
This matters for developers managing AI sub-agents. It offers zero-token-cost visibility into who is waiting on whom and who is blocked. That could help debug multi-agent workflows. Still, the project is mostly for fun.
The project is made by a bored dev and his robots. A built-in demo features a fictional team. The creator accepts coffee donations. The tone is playful rather than enterprise-serious.
Because it is open source and free, adoption depends on community interest. As AI agents become more common, such playful monitoring interfaces may inspire more practical dashboards. For now, Agent Standup remains a lightweight, experimental way to watch Claude Code sessions.
