agent-standup

Agent Standup

See which of your agents is working, stuck, waiting on another one, or burning tokens — across all your Claude Code, Codex, Mistral Vibe and Cowork sessions, at zero token cost. Agent Standup turns the session files Claude Code, Codex and Mistral Vibe already write into a live team call (and a board of every session): it only reads local files and never calls Claude unless the call is held in French, Spanish or Brazilian Portuguese, or you turn on the jokes.

No editor? Try it in 30 seconds from any terminal (Node 18+), then ⚙ → Source → Demo if you have no session yet:

npx @double_geste/agent-standup@latest --open

Your Claude Code, Codex or Mistral Vibe session as a live team call. The lead gives orders, sub-agents join the call, work at their desks, share their screen and report back — with voices, expressions and a view of the real work as it happens.

Works with: Claude Code · Codex · Mistral Vibe · Cowork — in VS Code, Cursor, Windsurf, any terminal, or as a desktop app.

Just installed the extension? Click the Agent Standup icon in the Activity Bar (or run Agent Standup: Open live call): the call starts at once, silent, and your first click in it turns the sound on. No session yet: press ▶ Watch the demo.

→ Website & download: https://agent-standup.com

🌐 Français · 繁體中文 · 简体中文 · Español · Deutsch · Dansk · Português · 日本語 · 한국어

A fictional team fixing a slow login page — the built-in demo (⚙ → Source → Demo). Watch the full 1-minute video.

New: Agent Standup is on YouTube. The first video is the very session that made it, replayed in the app with its own voices (2 min 51). Subscribe to the channel.

Halloween pack (optional)

Eight characters drawn by hand by Fabien Coupas, each in its own room: a pumpkin leads the call, you are a vampire, and a ghost, a cat, a bat, a wolf, a witch and a crowd of skeletons do the work.

It is entirely optional. Agent Standup is and stays free and MIT licensed, and nothing in it is locked behind the pack: every feature and the default characters stay as they are, and the vampire is free for everyone. Buying the pack (2 €, or more if you like) is simply a way to support the artist who drew it and the developer who makes the app; the money is shared between them.

See the pack, or in the app: ⚙ → Characters → Halloween pack. The pack's pictures are not part of the open-source code: they remain the artist's work.

Unofficial community project, not affiliated with or endorsed by Anthropic, OpenAI or Mistral AI. "Claude" is a trademark of Anthropic; "Codex" is a trademark of OpenAI; "Mistral" is a trademark of Mistral AI.

What you get

• Token usage per agent — a ⚡ counter in the header, on each tile, on each line of the agents list and in the camera of the agent who shares its screen shows how many tokens each agent used (input + output + cache writes; cache reads shown separately in the tooltip), counted once per message as recorded by Claude Code. Handy to see which sub-agent is the expensive one. When a session mixes models, each count says which model spent it (“⚡ 391 · Haiku”). It is read from the session files, not an estimate of your bill.
• Stuck or quiet agents — in the live call, a tile pulses amber with ⏳ when a command has been running for over a minute, shows 💤 when a working agent has gone quiet for 3 minutes, a red ✗ counter for failed tool calls, and a ♻ count of repeated work (a file read again with no change, the same search twice, a failing command run again). Hover any of these badges (or tap it) to see the calls behind it. The browser tab title shows how many agents are working.
• Built-in demo — no session yet? ⚙ → Source → Demo plays a fictional team (no data of yours).
• Live call — follows your newest session (Claude Code, Codex or Mistral Vibe) in real time. Each agent is a small character at its desk; the one who speaks gets the spotlight and keeps it until someone else talks. Agents never talk over each other: when several have something to say at once, they take turns, so every line is heard.
• Real voices, lip-sync — lines are spoken aloud by a small speech model on your machine (a one-time install); mouths follow the actual volume of the voice. Until then, and if you prefer it, the agents babble (see Voices).
• Live screen sharing with the real output — when an agent works, its screen fills the stage (presenter's camera in the corner, "sharing" badge): a terminal with the command history and the actual output of each command, search results and tool calls; or an editor with the file it reads (real content and line numbers) or the exact diff it applies. See What you see for each agent below.
• Agents come and go — a sub-agent joins when it is given a task and leaves the call 2 minutes (playback time) after its report, with a dissolve animation. It comes back if it gets a new task.
• Sub-agent tree — the full hierarchy (you → the lead → sub-agents → their sub-agents) with status, tool counts and the current action.
• Click any agent — its prompt, tool usage and reports.
• Replay — pick any past session, scrub through it, see a timeline of who worked when, and get a recap (agents, tool calls, commands, files changed, busiest agents).
• Video export — ⚙ → Tools → 🎞 Export video records the call at 1280×720 with voices and subtitles, as an MP4 (H.264 + AAC) when the browser can record it, and as a .webm when it cannot. X, phones and most chat apps refuse .webm.
• Clip to share — ⚙ → Tools → 🎬 Clip to share (25 s) records about 25 seconds, saved as agent-standup-clip.mp4 (.webm when the browser cannot record MP4). In the middle of a replay it starts at the current position; otherwise one spoken line before the first error or loop of the session, and failing that at the moment the lead first hands out work. The line being said when time is up is finished rather than cut, so the length varies by a few seconds. The clip carries the site address (atoll92.github.io/agent-standup) in the top right corner and ends with a short end card; the full export has neither.
• Hide code & file names in videos — ⚙ → Tools, off by default, for the full export and the clip. Commands, their output, file contents, diffs and search results on the shared screen are drawn as grey bars of the same length; file names are removed from the window title and from the activity tags on tiles (the tool's name, e.g. Bash or Read, stays); the session title is replaced by "Agent Standup"; pictures (screenshots, images) become an empty frame. What the agents say is not masked: speech bubbles, subtitles and voices are kept as they are, so check the clip before posting if agents quote code or names aloud.
• In French, Spanish or Portuguese too — the call can be held in French, Spanish or Brazilian Portuguese: agent text is translated on the fly and read by voices of that language (see below).
• Office music — optional live web radio (FIP and its themed stations) that fades down smoothly while an agent talks and fades back up when the room goes quiet.
• Office ambience, optional (off by default) and fully synthesised: keyboards that type when agents run commands, clicks when they read, a coffee machine, a phone buzzing, a far-away door… (⚙ → All settings → Office ambience). Plus small reactions and jokes between agents ("episode mode").

Claude Code plugin

You can also open the call from inside Claude Code. The plugin adds two skills: /agent-standup:open starts the app and opens the call in your browser, and /agent-standup:status says in one line who is working and who needs you.

/plugin install agent-standup --marketplace Atoll92/agent-standup

On Claude Code before 2.1.275: /plugin marketplace add Atoll92/agent-standup, then /plugin install agent-standup@double-geste. The plugin runs the same npm package as the command line, pinned to an exact version.

A line above your prompt. A second plugin, agent-standup-band, shows above the Claude Code prompt which other session waits on a question or an approval, and a short toast when one starts waiting or finishes its turn, in the terminal and in the Claude desktop app's Code tab (Claude Code 2.1.294 or later). It reads only ~/.claude/sessions, needs no Node.js and sends nothing:

/plugin install agent-standup-band --marketplace Atoll92/agent-standup

Codex sessions too

Sessions of OpenAI’s Codex (the desktop app and the codex command) appear next to your Claude Code sessions: in the session list (project “Codex”), on the board, and in the live call when one of them is the session you last typed in. Nothing to set up: the app reads the files Codex already writes in ~/.codex/sessions (or $CODEX_HOME), and never writes there.

• What is shown: what you typed, what the agent says, each command with its real output and whether it failed, files it created or changed, web searches, MCP calls, the questions it asks you, its reasoning summaries as inner thoughts, token counts, and the end of each turn.
• Compressed sessions too. Recent versions of Codex compress the sessions you are no longer in (rollout-….jsonl.zst). They are listed and replayed like the others, but reading them needs Node 22.15 or later: from the command line that is your own node; the desktop app has it; in VS Code, Cursor or Windsurf it is the editor's own runtime, so an older editor lists only the sessions Codex has not compressed.
• Codex sub-agents join their lead's call, with their own names, commands, results, token counts and final reports, and their own sub-agents too. They are not listed as sessions of their own. The short reviews Codex runs on its own before a command are not shown as sessions.