A fully managed OpenClaw instance that works with your files, tools, and workflows. No GPUs to provision. No API keys to juggle. Just sign up and start building.
Enter your phone number during signup — and your agent texts you. No app downloads. No bot invites. No developer console to wrestle with.
Just a real text message from your agent saying "Hey, I'm ready."
Browse the web, run code, manage files, schedule tasks. Your agent meets you on Telegram, Discord, WhatsApp, or the Web UI.
Your instance is live the moment you sign up. No GPU provisioning, no config files, no environment variables.
Browser automation, terminal access, file management, web search, code execution, messaging integrations, cron jobs, and more.
Telegram, Discord, WhatsApp, Web UI. Plus 300 SMS messages/month. Your agent follows you.
Monthly subscription. No per-token surprises. No metered API costs. What you see is what you pay every month.
Your agent remembers context across sessions. Skills, preferences, project details — it builds on what it already knows.
Schedule recurring tasks with cron jobs. Monitoring, data pipelines, periodic reports — running on their own.
No infrastructure knowledge required. We handle everything.
Pick a plan. Straightforward monthly billing, no surprises.
Add your messaging accounts. Telegram, Discord, WhatsApp, SMS — your agent is instantly reachable.
From anywhere. Any time. Your agent is always on, always working.
Same agent. Same tools. Only the token ceiling changes.
A fully featured agent. Predictable pricing. No surprises.
Everything Starter has — only faster, louder, and without limits.
7M words — most users never come close.
Everything you need to know before getting started.
Your own fully managed OpenClaw AI agent instance. It can browse the web, run terminal commands, manage files, execute code, integrate with messaging platforms, schedule recurring tasks, and more. Everything is hosted and maintained by us — you just interact with your agent.
The Starter plan ($49.99/mo) includes 10 million tokens per month and 300 SMS messages. The Unlimited plan ($99.99/mo) has no token cap, 1,500 SMS messages, priority inference resources, and higher rate limits. Both plans include the same agent, the same 23 tools, and the same messaging channels.
A token is roughly three-quarters of a word. 10 million tokens is about 7.5 million words of input and output combined — that's roughly 15,000 detailed conversations. Most individual users never come close. If you're running heavy automation or building something that talks to your agent all day, the Unlimited plan is your bet.
Your agent stops processing new requests at the end of your allotment. It doesn't overcharge — you're simply paused until the next billing cycle. You can upgrade to Unlimited anytime mid-cycle and pick right back up.
Telegram, Discord, WhatsApp, and our Web UI are all available on every plan. Starter includes 300 SMS messages per month, Unlimited includes 1,500. We'll help you set everything up during onboarding.
Your instance runs on Qwen3.6-27B, hosted on our infrastructure. You don't need to worry about model selection or inference setup — it's all handled.
Your agent's memory, skills, and session data belong to you. You can export your data at any time. If you cancel, we'll securely delete your instance and all associated data within 30 days.
Yes. No contracts or commitments. Upgrade, downgrade, or cancel anytime. Your instance stays active until the end of your current billing period.
Self-hosting means GPU infrastructure, API key management, model serving, ongoing maintenance, and unpredictable costs. With ShrimpHost, everything's handled. One flat rate — focus on using your agent, not maintaining it.