Skip to content
3 workers active right now

Manage tasks,
not agents.

An open source Mac app that autonomously manages AI coding agents.

Mando command center: task pipeline stats, a 56-day merge-activity heatmap, and a per-project task list where each task runs on its own coding agent, a mix of Claude Code, Codex, and GLM workers

Captain

The task list that
does the work

Describe what you want built. Captain assigns a Claude Code worker, gives it a clean worktree, and watches the session. When the PR lands, you review it. If it's wrong, send it back with a note.

A Mando task page: Captain's activity feed showing a Claude Code worker assigned on a clean worktree, then PR #142 opened and marked ready for review
Scout in Mando: a reading list of AI articles and videos with sources, search, and one-click research

Scout

Your reading list,
wired to your task list

Paste a URL. Mando pulls it down, extracts the article, and writes a summary. Worth building? Make it a task. Captain picks it up.

Telegram + CLI

Run it from your phone.
Or your terminal.

Captain pings you the moment a task needs a call: a question to answer, a PR to review. Approve, rework, or nudge a stuck worker right from the chat. Prefer the keyboard? The mando CLI drives the whole fleet from your terminal.

M
Mando
bot
Today
📎
Message
Terminal
$mando todo add "Add SAML 2.0 SSO for enterprise" -p atlas
Added item #142: Add SAML 2.0 SSO for enterprise
$mando tasks
Tasks (13 items)
queued=1 in-progress=3 awaiting-review=2 merged=7
$mando merge 142 -p atlas
{
"status": "merged",
"item_id": 142,
"pr": 142
}

FAQ