Deploy a Ghostfastest agent 0 → 1.
A dedicated Linux VM with frontier models pre-wired and Claude Code, Hermes, and OpenClaw pre-installed. One-click deploy to Discord, Telegram, Gmail, Lark, or WhatsApp.
Zero config
Sign up. VM provisioned, keys wired, agent runtime warm. About a minute end to end.
- ✓VM provisioned0.3s
- ✓Keys wired0.6s
- ✓Runtime warm0.9s
- ✓SSH ready1.2s
Every frontier model
Claude, GPT, Gemini, Groq, Bedrock. Routed through one key, metered at gateway rates.
Skills, ready
Claude Code, Hermes, and OpenClaw running out of the box. SSH in, run the one you want.
~/
├── claude-code
├── hermes
└── openclawDisconnect-safe
Close the laptop. The agent keeps working on its dedicated VM. Reconnect anytime.
- $exit # close the laptop
- vm-7f3a · still running
- agent · 23 tasks queued
- uptime · 4d 12h
From signup to shipping in under a minute.
Most agent platforms ask you to provision infra, set up keys, deploy code, and configure webhooks. We do all of that the second you sign up. Your job is to use it.
Sign up.
Connect a channel.
Send your first task.
Walk away.
Meet your users where they are.
Your agent is one process. The rails to Discord, Telegram, WhatsApp, Lark, and Gmail are ours. Toggle a channel on, paste in a token (or scan a QR), and your agent is live.
- Hot-swap channels without redeploying
- Bring your own bot tokens
- One agent across every surface
- Switch channels with one config flip
Start free.Scale when ready.
Free includes 1 agent and $1 in credit. Paid plans add more agents, bigger machines, and discounted model usage — metered through our gateway.
$1 welcome credit on signup
Cancel any time. Balance and usage visible in your dashboard.
Questions, answered.
Three runtimes: claude-code, hermes, and openclaw. Ready out of the box. Plus the usual toolchain: apt, uv, and npm. SSH in and install whatever else you need.