What Is Paperclip? The Open-Source AI Agent Orchestration Platform Explained

What Is Paperclip?

Paperclip is an open-source AI agent orchestration platform built on a Node.js server and React UI. It lets you manage teams of AI agents as structured organizations — complete with org charts, budgets, governance rules, and audit trails. Instead of treating agents as isolated chatbots, Paperclip treats them as employees inside a company you design.

The project is fully open-source and available on GitHub, with a hosted dashboard at paperclip.ing. If you have been following the rapid evolution of agentic AI in 2026, Paperclip represents the next logical step: moving from individual agent capability to organizational agent management.

At Codeloop, we have been deploying multi-agent systems for clients since early 2025. Paperclip addresses the exact gap we kept hitting — the need for a centralized layer that coordinates agents, enforces spending limits, and maintains a clear chain of accountability. After testing it across several internal projects, we can say it delivers on its core promise.

How Paperclip Works

The core idea is straightforward: you bring the agents, Paperclip brings the structure. Your agents can be anything — Claude Code agent teams, OpenClaw bots, Python scripts, shell commands, or HTTP webhooks. Paperclip does not care what powers the agent. It cares about what the agent is responsible for, who it reports to, and how much it is allowed to spend.

On the surface, the dashboard looks like a project management tool — tasks, statuses, timelines. But underneath, Paperclip maintains a full organizational graph: reporting lines, role definitions, budget allocations per agent, and governance policies that control what each agent can and cannot do autonomously.

Here is how the workflow typically looks:

Key Features of Paperclip

What separates Paperclip from a basic task runner or simple agent launcher is the depth of its organizational features. Here is what you get out of the box:

• Organizational Structure — Define hierarchies, roles, and reporting lines. Agents know who they report to and who reports to them. A manager agent can delegate tasks to subordinate agents and aggregate their outputs.
• Budget Management — Set per-agent budgets and per-tool-call pricing. Paperclip tracks spending in real time and halts agents that exceed their allocation. If you have struggled with keeping token costs under control, this feature alone is worth the setup.
• Audit Trail — Every instruction, every response, every tool call is recorded. You get a complete, searchable history of what each agent did, why it did it, and what it cost. This is critical for compliance and debugging.
• Governance — Configure approval gates for high-stakes operations. Certain actions can require human sign-off or manager-agent approval before proceeding. All configuration is versioned with full rollback support.
• Multi-Company Support — One Paperclip installation can manage unlimited companies. Each company gets its own org chart, budget pool, and governance rules. Ideal for agencies or consultancies managing AI operations for multiple clients.

Getting Started with Paperclip

Paperclip ships with a one-command onboarding process. If you have Node.js installed, getting started takes under five minutes:

This command scaffolds a default company, creates your first org chart, and opens the React dashboard in your browser. From there, you can register agents, define roles, and assign your first set of goals.

The onboarding wizard walks you through connecting your first agent. If you are already running OpenClaw bots, you can point your existing OpenClaw installation at Paperclip's coordination layer. For agents that communicate via the Model Context Protocol, Paperclip supports MCP-compatible endpoints natively.

Paperclip vs Individual Agents

The best way to understand Paperclip's role is this analogy: if OpenClaw is an employee, Paperclip is the entire company.

An individual agent framework like OpenClaw (or its open-source codebase on GitHub) excels at making a single agent powerful. It connects to LLMs, messaging platforms, and tools. It handles context, memory, and tool execution for one agent at a time.

Paperclip operates at a higher layer. It does not replace your agents — it organizes them. Here is what each layer handles:

You do not choose between Paperclip and an agent framework. You use both. Paperclip sits on top and manages whatever agents you already have. If you are evaluating which agent framework to use underneath Paperclip, our 2026 comparison of LangChain, CrewAI, AutoGen, and LangGraph can help you decide.

Real Use Cases for Paperclip

Based on our experience deploying agent systems — and the patterns we have seen across real-world AI agent case studies — here are the three use cases where Paperclip delivers the most value:

When Should You Use Paperclip?

Not every project needs an orchestration layer. If you are running a single agent that answers questions or performs one type of task, you probably do not need Paperclip — just use OpenClaw or Claude Code directly.

Paperclip becomes essential when you need coordination, not just capability. Specifically, you should consider it when:

• You are running three or more agents that need to share context or hand off tasks to each other
• You need budget controls to prevent a runaway agent from burning through your API credits overnight
• You require audit trails for compliance, debugging, or client reporting
• You want governance gates so that certain decisions still require human approval
• You are building toward a zero-human or low-human operation where agents handle end-to-end workflows

The inflection point we have observed at Codeloop is around the three-agent mark. Below that, a simple script or cron job can handle coordination. Above that, you start losing track of who did what, costs spiral, and debugging becomes painful. Paperclip solves this cleanly.

Architecture at a Glance

For the technically curious, Paperclip's architecture is deliberately simple:

• Node.js Server — Handles the orchestration logic, agent registry, budget enforcement, and governance rules. Exposes a REST API for programmatic control.
• React Dashboard — The visual interface for managing companies, org charts, agents, tasks, and budgets. Runs in the browser, connects to the server via API.
• Agent Adapters — Lightweight connectors that let Paperclip communicate with any agent type. Built-in adapters exist for shell commands, HTTP webhooks, and common agent frameworks.
• Persistent Storage — All org definitions, task histories, audit logs, and configuration versions are stored locally. No external database required for the default setup.

The system is stateless at the orchestration layer — agents maintain their own context and memory. Paperclip only tracks the organizational metadata: who is assigned to what, budget consumption, task status, and approval state.

Limitations to Know About

Paperclip is still a young project, and transparency matters. Here are the current limitations we have encountered:

• Early-stage ecosystem — The community and plugin ecosystem are still forming. Expect to write custom adapters for niche agent types.
• No built-in LLM — Paperclip is an orchestrator, not an agent runtime. You need to bring your own agents and LLM keys.
• Learning curve for governance — The approval gate and versioned config system is powerful but takes time to configure correctly for complex workflows.

None of these are dealbreakers, but they are worth knowing before you commit to a full deployment. For most multi-agent setups, the value far outweighs the friction.

Need Help Building a Multi-Agent System?

At Codeloop, we specialize in designing, deploying, and managing AI agent systems — from single-agent automation to full multi-agent orchestration with Paperclip. Whether you need help choosing the right architecture, configuring governance rules, or integrating agents into your existing workflows, our AI agent services team has done it before.

We handle the technical complexity so you can focus on what your agents should accomplish.

Book a Free Consultation