AI harness hosting

Your AI harness,
on its own microVM.

Pick a harness, pick a size, and get a dedicated, hardware-isolated microVM with TLS, persistent storage, and model access included. Live in about three minutes.

No API keys to bring. No servers to patch. Delete any time.

Three harnesses, ready to run

Every harness comes preinstalled with model access through our gateway.

Paperclip

Autonomous agent company

Issues, agents, and a board that drives them. Ships with Claude Code, Codex, and OpenCode CLIs. Claim it once, then invite your team with Paperclip’s own user management.

QM

Multi-agent chat workbench

Skills, runs, and a first-party web UI. QM manages its own accounts — the first user to sign up becomes the owner.

Hermes

The Nous Research agent

Conversations, memory, scheduled jobs, and a workspace — all in Hermes’s full-featured web UI, persisted on your disk.

Up and running in three steps

1

Pick a harness

Paperclip, QM, or Hermes — preinstalled and configured, with model access through the built-in gateway. No API keys to manage.

2

Pick a size

Starter, Pro, or Power. Every plan is a dedicated microVM — hardware-isolated, with NVMe persistent storage that survives restarts.

3

Open the web UI

Your harness is live at your-name.harnesser.co with TLS in about three minutes. Keep it private, or make it public and let the harness manage its own users.

Simple, per-microVM pricing

Every plan is a dedicated microVM. Change or delete any time.

Starter

$49/mo
  • 2 vCPU boost
  • 4 GB RAM dedicated
  • 64 GB NVMe persistent disk
  • Dedicated microVM
  • TLS + model gateway included
Get started

Power

$199/mo
  • 8 vCPU boost
  • 16 GB RAM dedicated
  • 256 GB NVMe persistent disk
  • Dedicated microVM
  • TLS + model gateway included
Get started

Isolation you don't have to think about

Each harness runs in its own microVM with a hardware isolation boundary — not a shared container. Your disk, your users, your agent's workspace. Nothing is shared with other tenants.

Create your harness