Paperclip

Open-source control plane for running teams of AI agents like a company

Visit website →

About

Paperclip is an open-source control plane for organizing and running teams of AI agents the way you would run a company. Instead of managing one assistant, you define an org chart with roles, reporting lines, and goals, then assign work that inherits context from the objectives above it.

It connects to any agent runtime, whether Claude Code, Codex, CLI tools, or plain HTTP bots, and coordinates them from a single dashboard. Scheduled heartbeat runs keep work moving around the clock, per-agent budget caps throttle spending automatically, and approval workflows hold risky actions for a human.

Every action lands in an immutable audit log with rollback, and one deployment can isolate several separate organizations. Paperclip is MIT-licensed and self-hostable via a one-line installer or Node.js and pnpm, while a hosted plan runs ten euros a month with unlimited companies and no per-seat fees.

AI-assisted draft, human-reviewed before publishing — see how we choose & review tools.

Why we picked it

Paperclip stood out for the org-chart framing done literally — roles, reporting lines, and inherited goals instead of a flat list of prompts, and it plugs into whatever agent runtime you already use. Worth it for someone scaling past a single assistant who needs actual structure.