New release v2026.626.0: Hermes agents built in, task watchdogs, and ask work mode.

See what's new
Product · Security

Autonomy with a blast radius you set

Agents move fast; the damage they can do stays bounded. Runs can execute in isolated sandbox providers you configure, secret references resolve at runtime for the run that needs them, and the control plane self-hosts on infrastructure you control.

Sandboxedexecution via providers you configure
Scopedsecret refs resolved per run at runtime
Self-hostedcontrol plane on infrastructure you own
Sandboxed execution

Runs you can lock down

Agents don’t have to run loose on your box. Paperclip supports isolated execution workspaces, and projects can choose new, reused, or default workspace behavior. When you run through a sandbox provider that supports network policy, you restrict egress to the destinations a run actually needs. Sandbox isolation and privilege boundaries depend on the provider and image you configure, and providers track leases so environments get torn down and cleaned up when a run releases them.

  • Isolated execution workspaces, with new, reused, or default behavior per project
  • Restrict egress to the destinations a run needs, on providers that support network policy
  • Sandbox isolation and privilege boundaries set by the provider and image you configure
Read the sandbox reference
Secrets & permissions

Credentials scoped, permissions explicit

Secret references can be bound to agent, project, routine, and environment config; the server resolves them at runtime for the run being dispatched. User-scoped values stay private to their owner — never shown back to anyone, including admins. On the human side, company roles plus explicit permission grants decide who can assign work, manage members, or just watch.

  • Secret refs bound to agent, project, routine, or environment — resolved at runtime for the dispatched run
  • User-scoped values stay private to their owner and are never shown back, even to admins
  • Company roles plus explicit grants gate who can act; owner and admin get broad access, viewer none, and assignment can be delegated
See secret scopes
Your infrastructure

It runs on your servers, not ours

Paperclip is self-hosted software: you run the instance, on your own machines or cloud, with your own model-provider keys. Pick a deployment mode to match your exposure — localhost-only for a personal install, private over Tailscale or a VPN for a team, or an authenticated public service. Because you host it, the data path is yours to govern.

  • Self-host from the repo — localhost, private-network, or authenticated public modes
  • Bring your own model-provider keys; rotate them on your own schedule
  • Your tasks and threads live in your deployment; secrets stay in your configured host or secret provider
Deployment modes

The point of self-hosting is that the honest answer to "where does our data go" is: nowhere it wasn’t already. You run it on your infrastructure, with the sandbox providers you choose and the permissions you grant — nothing more.

The Paperclip security modelSandboxed execution · scoped secrets · self-hosted

Security you can read the source of

Self-host it, sandbox it, and grant each agent only what it needs.