24/7 autonomous AI agents in your own cloud VPS
LiveBridge rents you a dedicated cloud server running 1, 5, or 10 autonomous AI agents. They work around the clock on your tasks, remember context across sessions, and Claude usage is included in your tier — no LLM key to manage, no second bill.
See pricing → Sign in Open your workspace → Manage billing
How it works
- Pick a tier on the pricing page and pay once a month.
- We boot a dedicated DigitalOcean VPS in your name in about 90 seconds, pre-installed with the agent runtime, Postgres, and TLS.
- Sign in and start chatting — Claude is wired up automatically. Your tier includes a monthly LLM budget; we route the calls and pay Anthropic. No keys, no second bill.
- Define agents in plain English. Each one runs in the background, follows a schedule or fires on triggers, and writes results to a persistent memory store you can browse.
What the agents can do today
- Browse the web — navigate, fill forms, extract data, take screenshots.
- Read and write files on the VPS (and on your Mac for Developer tier).
- Run shell commands in an isolated sandbox.
- Talk to 50+ external services — Gmail, Calendar, GitHub, Notion, Linear, Slack, Discord, Telegram, Spotify, and more.
- Remember what they learned across restarts via local vector search over their own notes.
- Coordinate with each other — multiple agents on one VPS share memory and can hand work off.
- Send notifications when interesting things happen (email / Slack / push).
LLM usage included
Every tier includes a monthly Claude budget. We hold the API key, meter every request, and stop at your cap so you never get a surprise bill. Plus every new customer gets a $5 trial credit on signup — stacks on top of your monthly budget. Specifically:
- Starter $29 includes $5/mo of Claude usage (~250K Sonnet tokens with prompt caching).
- Pro $49 includes $15/mo (~800K Sonnet tokens).
- Enterprise $89 includes $40/mo (~2.5M Sonnet tokens).
- + $5 one-time trial credit on every new subscription.
- Hit the cap? Auto-routes to free Llama 3.3 until next month, or upgrade your tier.
- No keys in your dashboard, no second bill, no manual setup.
Tiers
- Starter $29/mo — dedicated VPS, 1 agent.
- Pro $49/mo — up to 5 agents on a larger VPS.
- Enterprise $89/mo — up to 10 agents, daily backups, priority support.
- Developer $199/mo — everything in Enterprise plus a Mac connector. Install the daemon on one Mac and your cloud agents can run commands, take screenshots, and drive Claude Code there directly.
- Developer Pool $399/mo — up to 5 connected Macs that your cloud agents can choreograph as one fleet.
We’re in private build today. Drop your email on the
pricing page and we’ll send you an
invite when the runtime is live in production.
Architecture
- One isolated DigitalOcean droplet per customer — no shared compute, no noisy neighbours.
- Caddy with auto-issued Let’s Encrypt TLS for every customer subdomain.
- Postgres 16 + ufw firewall + fail2ban out of the box.
- Source code never lives on customer machines — we ship pre- built Docker images from a private registry.
- Provisioning is fully automated and idempotent. If anything fails, we tear it down cleanly and refund.
Built on OpenClaw (MIT, sponsored by OpenAI / GitHub / NVIDIA / Vercel) and Anthropic’s Claude ecosystem.