Paperclip runs a team of AI agents like a company

Paperclip (opens in a new tab) treats a group of AI agents as a company. You set a goal, hire agents into an org chart (CEO, CTO, engineers, marketers), approve the CEO agent's strategy, and watch the work from a dashboard. It looks like a task manager. Underneath, it handles budgets, reporting lines and governance.

The parts that stood out:

• Bring your own agent. Claude Code, Codex, Cursor, Gemini CLI, OpenCode, Pi and others can share one org chart. Per the README: "If it can receive a heartbeat, it's hired."
• Heartbeats. Agents wake on a schedule or when they're assigned a ticket, so they don't sit in a loop burning tokens.
• Hard budgets. Each agent gets a monthly spend limit and is paused when it hits it.
• You're the board. Hiring a new agent needs your approval by default, and every ticket keeps a full tool-call trace and an append-only audit log.

What makes it worth more than a bookmark is that the whole thing is on GitHub (opens in a new tab) under the MIT license. It's a Node.js server and React UI with an embedded Postgres, self-hosted, with no Paperclip account required:

npx paperclipai onboard --yes

That matters for teams. You can run it inside your own infrastructure, point it at your own Postgres, and write a custom adapter for an in-house agent instead of waiting for a vendor to support it. And when an agent does something odd, you can read the orchestration code instead of guessing at a black box. With an agent manager, that's what you actually need.

My caveat: a company of agents is only as good as its checks. Budgets and audit logs tell you what happened. Deciding whether the work was any good is still your job.