Paperclip: The Open-Source Agent Management Platform With 77,900+ GitHub Stars That Turns AI Agent Chaos Into an Organized Company

If you've ever had 20 Claude Code terminals open, lost track of which agent was doing what, or watched helplessly as a runaway loop burned through hundreds of dollars in tokens, you've experienced the pain that Paperclip was built to solve.

Paperclip isn't another chatbot wrapper or agent framework. It's a full orchestration layer that treats your AI agents like employees in a company—with org charts, job titles, budgets, approval workflows, and performance reviews. Think of it this way: if Claude Code is an employee, Paperclip is the company.

What Is Paperclip?

Paperclip is an open-source Node.js server and React UI that orchestrates a team of AI agents to run a business. You bring your own agents (Claude Code, Codex, Cursor, or any bot that can receive a heartbeat), assign goals, and track work and costs from one dashboard.

It looks like a task manager on the surface. Under the hood, it's a complete control plane with:

• Org charts for mixed human + agent teams
• Task management with atomic checkout and execution locks
• Budget enforcement that pauses agents when they hit limits
• Governance workflows with approval gates and audit trails
• Heartbeat execution for autonomous 24/7 operation

The tagline says it best: "If it can receive a heartbeat, it's hired."

The Four Pillars of Agent Orchestration

Paperclip is built around four core pillars that every autonomous organization needs:

1. Agentic Task Manager

Declare your intent. Agents work on tasks. You verify the output. Tasks carry full context—company goals, project objectives, parent issues—so agents always know why they're working on something, not just what.

Key features:

• Atomic task checkout prevents double-work
• First-class blocker dependencies
• Comments, documents, attachments, and work products
• Inbox state for review and approval

2. Org Chart for Agents

Agents have roles, titles, reporting lines, and permissions—just like human employees. You can build hierarchies where a CTO agent manages engineering agents, who in turn manage specialized coding agents.

The org chart supports:

• Mixed human + agent teams
• Delegation and specialization
• Governance: who can do what
• Scoped secrets and company boundaries

3. Agent Employee Training

Design, train, and evaluate your AI employees. Paperclip includes a Skill Studio for creating shared org-wide skills, evals for testing agent performance, and active learning loops for continuous improvement.

You can run performance reviews for agents, track quality metrics, and save test runs to compare different approaches.

4. Agentic OS

The infrastructure that makes everything run. This includes cross-provider runtime (any model, any agent), sandboxing, integrations, MCP servers, SSO, GRC, RBAC, cost controls, and data privacy.

Every conversation is traced. Every decision is explained. Full tool-call tracing and immutable audit logs give you complete visibility into what happened and why.

Real-World Example: Running an Autonomous SaaS Business

Let's say you want to build "the #1 AI note-taking app to $1M MRR." Here's how Paperclip makes that happen:

1. Define the goal: Create a company-level objective in Paperclip
2. Hire the team: Spin up CEO, CTO, engineering, design, and marketing agents (or humans)
3. Set budgets: Assign monthly token budgets to each agent ($500 for engineering, $200 for marketing, etc.)
4. Create tasks: Break the goal into projects and issues—"Build auth system," "Design landing page," "Write launch blog post"
5. Approve and run: Review strategy, set approval gates, hit go
6. Monitor: Watch agents work from the dashboard, approve key decisions, pause or redirect as needed

Agents wake on heartbeats, check their tasks, do work, and report back. If an engineering agent hits its $500 budget, it stops automatically. If a task needs approval, it queues for human review. Everything is tracked, auditable, and resumable.

Key Benefits

• Eliminate agent chaos: No more 20 open terminals with no visibility. Tasks are ticket-based, conversations are threaded, sessions persist across reboots.
• Prevent runaway costs: Token budgets with hard stops. When agents hit limits, they pause. No more surprise $1,000 bills.
• Maintain context: Context flows from task → project → company goal. Agents always know what they're doing and why.
• Enforce governance: Approval gates, audit logs, and rollback capabilities. Nothing ships without your sign-off.
• Run 24/7 autonomously: Heartbeats wake agents on schedules. They check work, act, and report—while you sleep.
• Manage from anywhere: React UI works on desktop and mobile. Approve hires, override strategy, pause agents from your phone.
• Multi-company isolation: One deployment can run many companies with complete data separation.

How It Works Under the Hood

Paperclip handles the hard orchestration details correctly:

• Atomic execution: Task checkout and budget enforcement are atomic—no double-work, no runaway spend
• Persistent agent state: Agents resume the same task context across heartbeats instead of restarting
• Runtime skill injection: Agents learn Paperclip workflows and project context at runtime, without retraining
• Governance with rollback: Approval gates are enforced, config changes are revisioned, bad changes can be rolled back
• Goal-aware execution: Tasks carry full goal ancestry so agents see the "why," not just a title
• Portable company templates: Export/import orgs, agents, and skills with secret scrubbing
• True multi-company isolation: Every entity is company-scoped for complete data separation

Getting Started

Paperclip is open source (MIT license) and self-hosted. Installation is a single command:

curl -fsSLO https://paperclip.ing/install.sh
bash install.sh

The installer ensures Node.js 20+ is available, installs a managed Paperclip CLI under ~/.paperclip/cli, and starts interactive onboarding. It can also install Paperclip as a background service on Linux and macOS.

No Paperclip account required. Your data stays on your infrastructure.

Frequently Asked Questions

Is Paperclip free to use?

Yes, Paperclip is completely open source under the MIT license. You self-host it on your own infrastructure. The only costs are the AI model API calls your agents make.

What AI agents does Paperclip support?

Paperclip supports any agent that can receive a heartbeat. Out-of-the-box adapters include Claude Code, Codex, CLI agents (Cursor, Gemini, bash), and HTTP/webhook bots. You can also write custom adapters for any agent runtime.

How does budget enforcement work?

You set monthly token budgets per agent (e.g., $500 for engineering, $200 for marketing). Paperclip tracks token usage and costs in real-time. When an agent hits its limit, it automatically pauses and cancels queued work. No runaway spending.

Can I use Paperclip with multiple companies?

Yes. Paperclip supports true multi-company isolation. One deployment can run many companies with complete data separation, separate audit trails, and independent org charts.

What's the difference between Paperclip and OpenClaw?

OpenClaw is a personal AI assistant—an employee. Paperclip is the company that manages many employees (agents). You can use OpenClaw agents inside Paperclip, or any other agent runtime. Paperclip provides the orchestration layer; OpenClaw provides the individual agent intelligence.

Do I need to keep my computer running?

No. Paperclip runs as a server with a heartbeat system. Agents wake on schedules, check work, and act autonomously. You can manage everything from your phone and only intervene when needed.

How does approval workflow work?

You define approval gates for sensitive operations (hiring agents, deploying code, spending above thresholds). When an agent needs approval, the task queues for human review. You approve, reject, or modify from the dashboard.