Skip to content
Get packs

Guide · Open source

How to use Paperclip: install, set up agents and track work

Updated · written with AI from the project's own documentation and checked against our catalogue data

Paperclip is an open-source app that uses a Node.js server and a React dashboard to coordinate a team of AI agents toward business goals. It is built for developers and small teams who already run several agents and want one place to assign goals, track work and costs, and keep an audit trail. This guide explains what you need, how to get the code, and how to use the core concepts once it is running.

Paperclippaperclipai/paperclip · ★ 110k · MIT

Quick start

  1. 1Check prerequisites. Install Node.js and Git on the machine where Paperclip will run.
  2. 2Clone the repository. Run git clone https://github.com/paperclipai/paperclip.git to download the source.
  3. 3Build and start. Follow the build instructions in the README and docs for the clone-and-build method.
  4. 4Connect an agent. Use the adapter overview in the docs to connect the agents you already run.
  5. 5Define a goal. Create an organization goal and link tasks or projects to it.
  6. 6Set budgets and approvals. Assign budgets and review gates before agents start autonomous work.
  7. 7Review output. Check diffs, screenshots, tests and run history before accepting results.

What Paperclip is and who it is for

Paperclip looks like a task manager, but underneath it adds org charts, budgets, governance, goal alignment and agent coordination. You bring your own agents, choose models and harnesses per agent, and keep tasks, skills, permissions and history in one dashboard.

The README says it suits developers already running several AI coding agents, small teams testing autonomous agents that need budgets and approval steps, and anyone who wants an audit trail of what each agent did and why. It is not a hosted product: the README points to a Paperclip Cloud waitlist rather than a live service.

  • Good fit: you run several agents and need one place to track them
  • Good fit: you want budgets and approval steps before agents act
  • Poor fit: you want a finished hosted product with no setup
  • Poor fit: you only need a single-agent chat tool

Install Paperclip from source

The install method listed for Paperclip is clone-and-build. The prerequisites are Node.js and Git. Start by cloning the repository from GitHub using the command below, then follow the build steps in the project's README and docs.

Because Paperclip is a self-hosted Node.js app, you are responsible for running the server on your own machine or infrastructure.

  • Install Node.js
  • Install Git
  • Clone the repository
git clone https://github.com/paperclipai/paperclip.git

Connect your agents

Paperclip does not ship its own agents. The README lists supported agents including OpenClaw, Claude Code, Codex, Cursor and Cloud, Gemini CLI, OpenCode, Pi, Hermes, Grok Build and Kimi Code. Custom processes, HTTP endpoints and external adapter packages can also join the roster.

Setup details for each adapter are in the adapter overview in the Paperclip docs. Check that page for the steps and capabilities of the agent you plan to use.

  • Named integrations are listed in the README's Works with section
  • Custom processes and HTTP endpoints can extend the roster
  • The adapter overview at docs.paperclip.ing covers setup

Set goals, org chart and budgets

Paperclip works from goals rather than individual tasks. You define a goal, add agents to an org chart with roles, and set budgets. The README's example goal is building an AI note-taking app to a revenue target, with a CEO, CTO, engineers and marketers as the team.

Tasks and projects can be linked to organization goals, and agents receive the goal context behind their work. Delegation flows up and down the org chart. Costs are tracked from the same dashboard.

  • Define the goal the organization works toward
  • Hire agents into roles on the org chart
  • Approve strategy, set budgets and start the run
  • Monitor progress and costs from the dashboard

Review agent work and keep an audit trail

The README stresses verification and oversight. Paperclip describes tasks, approvals and review gates, and the ability to verify output from diffs, screenshots and tests. Run history and opt-in tracing are listed as part of the infrastructure.

Agents can wake for assigned work, follow-up messages or configured schedules, so plan who reviews output before you let them run unattended. Teams without a review process are a poor match.

  • Use approvals and review gates before work is accepted
  • Verify output from diffs, screenshots and tests
  • Check run history for what each agent did and why

What people use it for

  • Tracking several Claude Code or Codex terminals in one dashboard
  • Running a small set of agents toward a shared business goal with budgets
  • Keeping an audit trail of what each agent did and why
  • Assigning follow-up work to agents on a schedule

Frequently asked questions

Is Paperclip free?

The source lists the project under the MIT licence and the repository is open source. The README does not describe hosted pricing; it points to a Paperclip Cloud waitlist instead.

What do I need to install Paperclip?

The listed prerequisites are Node.js and Git. You clone the repository and build it yourself, so you need to be comfortable self-hosting a Node.js app.

Does Paperclip have a hosted version?

The README links to a Paperclip Cloud waitlist rather than a live hosted service. If you need a finished product with no setup, Paperclip is not the right fit yet.

Which AI agents does Paperclip work with?

The README lists OpenClaw, Claude Code, Codex, Cursor, Gemini CLI, OpenCode, Pi, Hermes, Grok Build and Kimi Code. Custom processes, HTTP endpoints and external adapter packages can also be added.

How does Paperclip compare to CrewAI, AutoGen, LangGraph or OpenHands?

Those are comparable projects for coordinating agents. Paperclip's distinguishing focus, per its README, is the business layer: org charts, budgets, governance and an audit trail on top of agent coordination. Compare the features you need before choosing.

Can Paperclip manage a single chat agent?

Not really. The README says Paperclip is built for coordinating multiple agents, so a simple single-agent chat tool will be a better match for that job.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

Paperclip suits developers and small teams who already run several AI agents and want budgets, org charts, approvals and an audit trail in one self-hosted dashboard. It is a poor choice if you want a finished hosted product, a single-agent chat tool, or you have no plan for reviewing what agents produce.

Install options & downloads