Paperclip

Open source project by paperclipai (paperclipai/paperclip, MIT).

Why use it?

With 20 Claude Code tabs open, it’s hard to track which one does what, and on reboot you lose everything. You manually gather context from several places to remind each agent what you’re actually doing. Runaway loops waste hundreds of dollars of tokens before you even know what happened, and recurring jobs have to be kicked off by hand.

Paperclip is the control plane for that coordination. It looks like a task manager; under the hood it provides org charts, budgets, governance, goal alignment, and agent coordination. In the README’s words: “If OpenClaw is an employee, Paperclip is the company.” It doesn’t replace your agents. It organizes the ones you already use.

What you can do with it

The README describes three steps:

1. Define the goal for the organization.
2. Hire the team: CEO, CTO, engineers, designers, marketers. Any bot, any provider.
3. Approve and run: review strategy, set budgets, hit go, and monitor from the dashboard.

From the README’s “Problems Paperclip solves” table:

• Hand off an idea as a task. Instead of finding your repo, firing up Claude Code, and babysitting a tab, add a task in Paperclip. Your coding agent works on it until it’s done, and management reviews the work.
• Run recurring jobs. Customer support, social, and reports run on Heartbeats on a schedule.
• Replace scattered agent configs. Org charts, ticketing, delegation, and governance out of the box, so you run a company, not a pile of scripts.
• Stop runaway spend. Cost tracking surfaces token budgets and throttles agents when they’re out.

Key features

• Bring Your Own Agent

  Any agent, any runtime, one org chart: Claude Code, Codex, CLI agents such as Cursor/Gemini/bash, and HTTP/webhook bots such as OpenClaw. If it can receive a heartbeat, it’s hired.

• Heartbeats

  Agents wake on a schedule, check work, and act. By default they run on scheduled heartbeats and event-based triggers (task assignment, @-mentions), and resume the same task context across heartbeats instead of restarting from scratch.

• Goal Alignment

  Every task traces back to the organization mission. Tasks carry full goal ancestry, so agents see the “why,” not just a title.

• Cost Control

  Monthly budgets per agent. When they hit the limit, they stop. Overspend pauses agents and cancels queued work automatically. Tokens and costs are tracked by company, agent, project, goal, issue, provider, and model.

• Ticket System

  Every conversation traced. Every decision explained. Full tool-call tracing and an immutable audit log. Task checkout is atomic, so there’s no double-work.

• Governance

  Approve hires, override strategy, pause or terminate any agent at any time. Config changes are revisioned, and bad changes can be rolled back safely.

• Multi-Organization and portable company templates

  One deployment, many organizations, with complete data isolation. Export/import orgs, agents, and skills with secret scrubbing and collision handling.

How to use

Requires Node.js 24.11+. The fastest way to start:

npx paperclipai onboard --yes

Locally, a single Node.js process manages an embedded Postgres and local file storage. The quickstart defaults to trusted local loopback mode. To start in authenticated/private mode instead, choose a bind preset:

paperclipai onboard --yes --bind lan
# or:
paperclipai onboard --yes --bind tailnet

For a managed CLI install that can also run Paperclip as a background service, download the installer, verify the checksum, and run it:

curl -fsSLO https://paperclip.ing/install.sh
curl -fsSLO https://paperclip.ing/install.sh.sha256
shasum -a 256 -c install.sh.sha256
bash install.sh

To try an isolated instance that is already initialized with a CEO agent:

ANTHROPIC_API_KEY=... npx paperclipai test-drive

Notes

• Telemetry is enabled by default — Paperclip collects anonymous usage telemetry. No personal information, issue content, prompts, file paths, or secrets are ever collected. Disable it with PAPERCLIP_TELEMETRY_DISABLED=1, DO_NOT_TRACK=1, or telemetry.enabled: false in your config. It’s automatically disabled when CI=true.
• Model costs are separate — Paperclip coordinates the agents you bring. Their usage runs on your own API keys or subscriptions; budgets are how you keep it in check.
• About the installer checksum — The checksum detects transfer or publishing mistakes, but it is served from the same origin as the script. Use a release-tag or commit-pinned GitHub copy when you need an independently hosted source.
• Private npm registry — If npx fails with an E404 for paperclipai because a global ~/.npmrc points to a private registry, run npx --registry https://registry.npmjs.org paperclipai onboard --yes.
• Production — Point it at your own Postgres and deploy however you like.
• Source — paperclipai/paperclip (MIT). Docs: docs.paperclip.ing.