Welcome to the Paperclip hosted beta

Build a team of AI agents, delegate work, and review results in one place. During this invitation-only beta, we handle hosting and maintenance so your agents can work even when your computer is off.

Get set up

  1. Sign in. Open your invitation link and sign in with Google, GitHub, or email. Use the email address that received your invitation.
  2. Create your organization. Give your organization a name.
  3. Name your first agent. Your first agent will help you plan work and build your team. Give it a name, then select Next.
  4. Connect a model. Choose Claude or OpenAI, then follow the prompts to connect your subscription or an API key.
  5. Finish setup. When the Review screen says your agent is ready, select Get started.

Start delegating work

Your first agent will ask about the work you want to delegate. Let it interview you and propose a plan, or select I have a task in mind.

Bring a task, an ongoing workflow, or a larger project. Explain the outcome you want and share relevant context. You can add agents and connect the tools they need as you build your team.

Follow progress in the task, answer questions, and review results. When your agent requests approval, review the proposed action before approving it or requesting changes.

See the guides to managing tasks and tracking agent activity.

Help shape the future of Paperclip

Tell us what’s working—and especially what isn’t. When something gets stuck, tell us what you expected and what happened. Your feedback will help us decide what to improve next.

Ask questions, share feedback, and get support in the private beta Discord or through in-product chat. If you can’t sign in, email support@paperclip.ing.

Thanks for helping us build Paperclip.

Frequently asked questions

What carries over when I connect Claude or OpenAI?

Connecting Claude or OpenAI lets Paperclip use your account to run agents. It doesn’t import your existing conversations, local files, skills, or MCP connections. Add the context and connections your agents need.

Will my agents keep working when I close my browser?

Yes. Agents run in a hosted environment, so your computer can be off or offline while they work. Tasks can still pause for input, approvals, or usage limits.

Is the hosted beta free?

Yes. We won’t charge for hosting or agent environments during the beta. Your model provider’s charges and usage limits still apply.

How much model usage should I expect?

Usage depends on the model, task, context, and how often agents run. Scheduled work also uses your provider allowance or API credits.

Check your provider account for billing and subscription limits. Paperclip’s spend totals don’t show your remaining subscription allowance.

See Costs and Budgets and Heartbeats and Routines to understand costs and when agents run.

Can I connect my tools and repositories?

Use Connectors to see the available connections and authorize access. Set them up in your hosted organization; connections from your local installation don’t carry over.

Can I import an existing company?

Company import isn’t available in this beta. Create a new organization and set up its agents and connections.

Does the existing documentation apply to the hosted beta?

Use the documentation for product capabilities and workflows. You can skip self-hosted installation and runtime setup—we manage those for you.