TECHNOLOGY
Agent Standup turns Claude Code sessions into a live stand-up call, free and open source
A free, open-source tool called Agent Standup shows what each Claude Code agent is doing in a stand-up call format. It reads local session files and runs in VS Code, Cursor, a terminal or a desktop app.
Image: Agent Standup
Agent Standup is a free, open-source tool that shows your Claude Code session as a live team meeting. The idea is to see at a glance which AI agents are working, waiting or blocked.
Here is how it works. The lead agent gives orders, and sub-agents join the call when they get a task. They share their screen, report back, and leave when they are done. The project's own page calls it "a bit silly" but "weirdly useful."
The tool reads the session files that Claude Code already writes on your machine. Because of that, it says it works wherever you run Claude Code, with nothing to configure and nothing sent anywhere. It also says checking status costs zero tokens. Tokens are the units AI models use to measure and charge for text.
There are several ways to run it. A VS Code extension (version 1.24.1) adds a panel, a side bar and status bar counts that turn red when an agent may need you. Cursor, Windsurf and VSCodium users can install the same extension from a .vsix file. In any terminal with Node 18 or newer, the command npx @double_geste/agent-standup --open opens the call in a browser.
There is also a desktop app with its own window and a menu bar indicator. The macOS builds are 94 MB for Apple Silicon and 100 MB for Intel. They are not signed by Apple, so on first launch you right-click the app and choose Open. Windows and Linux builds are listed as experimental. For Claude Code and tmux, a one-line status indicator is available after a global npm install. JetBrains has no native plugin yet, but you can add an External Tool that runs the terminal command.
Each agent gets one status: working, waiting on another agent, needs you, stalled or done. The list is sorted by urgency, and a Sessions board covers all your projects. A context gauge, shown as something like "ctx 62%", warns you before auto-compaction, and the tool shows tokens used per agent.
You can replay past sessions with highlights such as the first error, the longest wait and loops. It also offers a Markdown recap, a tree of who asked whom, and video export. Voices come with lip-sync in cartoon, Animal Crossing or Simlish styles, with optional office sounds and FIP radio. Users can add their own pixel-art characters by dropping PNG files in a folder. The interface comes in eight languages.
To use it, you must have run Claude Code on the machine at least once. Voices work best on a Mac. Support for other agent tools, including Codex CLI and Gemini CLI, is on the roadmap.