Paperclip: An Open-Source Platform for Managing AI Agent Teams

When working on projects with AI agents, the biggest issue isn’t whether a single agent can get the job done, but rather how to manage a group of agents. Imagine you have 20 Claude Code terminals running simultaneously, each doing something different; it’s practically impossible to keep track of what everyone is up to. That’s exactly the problem Paperclip is designed to solve.

What is Paperclip?

In a nutshell: It allows you to manage a team of AI agents just like you would manage a company.

It’s an open-source Node.js server with a React interface that’s used to orchestrate and manage multiple AI agents. You can assign roles to agents (such as CEO, CTO, engineer, designer), set goals and budgets, and then monitor the work of all agents from a single dashboard.

It has 69,000 stars on GitHub and is licensed under the MIT License.

Key Features

1. Access for Any Agent

Paperclip supports a variety of agents, including OpenClaw, Claude Code, Codex, Cursor, Bash, HTTP, and more. As long as an agent can send a “heartbeat” signal, it can be managed.

“If it can send a heartbeat signal, it’s considered ‘hired’.”

2. Organizational Structure

You can define roles and reporting relationships for agents, such as CEO, CTO, engineer, designer, and marketer. Each agent has a specific position and job description.

3. Goal Alignment

Every task is aligned with the company’s overall goals. Agents know what they’re working on and why.

4. Budget Control

Each agent has a monthly budget limit. Once the limit is reached, the agent stops working to prevent unexpected expenses, which is especially important for agents that are billed via APIs.

5. Heartbeat Mechanism

Agents wake up according to a preset schedule, check their work status, and then execute tasks. Task assignments are distributed based on the organizational structure.

6. Ticketing System

Every conversation and decision is recorded, with complete traceability of tool calls and immutable audit logs.

7. Multi-Company Management

A single deployment can manage multiple “companies” with completely isolated data. One control panel manages your entire group of agents.

8. Mobile Support

You can monitor and manage your agent team from your mobile device.

Installation and Usage

The simplest way to get started is:

npx paperclipai onboard --yes

This will start the local loop mode for quick setup. For local network access:

npx paperclipai onboard --yes --bind lan

For manual installation:

git clone https://github.com/paperclipai/paperclip.git
cd paperclip
pnpm install
pnpm dev

After installation, visit http://localhost:3100. It uses an embedded PostgreSQL database and doesn’t require any additional configuration.

Paperclip requires Node.js version 20 or later and pnpm version 9.15 or later.

Use Cases

• Multi-Agent Project Management: Run multiple AI agents simultaneously to perform different tasks, all from a single interface, without having to switch between 20 terminals.
• Automated Business Processes: Set goals and roles to enable the agent team to collaborate automatically on complex tasks (e.g., one agent writes code, another tests, and another deploys).
• Cost Monitoring: Track each agent’s API usage in real time and set budget limits to avoid unexpected expenses.
• Team Collaboration: Multiple people can share a Paperclip instance and manage their own groups of agents.

Summary

Paperclip elevates AI agent management from simply “running a bunch of terminals” to managing a “virtual company.” If you use multiple AI tools for your projects or want your agents to collaborate automatically, Paperclip is worth a close look.

GitHub: github.com/paperclipai/paperclip