Managed AI agent hosting · OpenClaw & Hermes

Hosted AI agents
that do real work.

Launch an always-on, isolated agent and put it on QA, research, outreach and ops — message it from your browser or any chat app, no servers to run.

Start hosted OpenClaw → See how it works

From $6.90/mo · setup in ~2 min · bring your own model keys

app.lobsterland.ing
Lobsterland built-in chat — your crew, live (dark theme) Lobsterland built-in chat — your crew, live (light theme)
~2 minto a running instance
99.9%uptime target
500+models, your own keys
No public IPisolated by default
Rated 5/5 on F6S & G2

New — two runtimes, one platform

Now hosting Nous Hermes alongside OpenClaw.

Pick the runtime that fits the job and run it with zero ops — same signup, same dashboard, same billing. Want the full operating surface? OpenClaw. Want a focused chat-and-channels agent? Hermes. Run either or both on one account, and switch experiments without re-onboarding.

Compare the two runtimes →
Full surface OpenClaw

Dashboard chat, workspace files, skills, cron, multiple agents, Hosted Browser, Paperclip, usage, and messaging channels.

Explore OpenClaw hosting → Chat-first runtime Hermes

Managed Nous Hermes hosting for browser chat plus Telegram and Slack with allowlist access control — online in minutes, no servers to run.

Explore Hermes hosting →

The product

Run the whole crew from one dashboard.

No SSH, no YAML, no Docker. Launch instances, configure agents, edit the workspace, manage skills and providers, and watch usage — all from the browser.

app.lobsterland.ing/app/instances
Lobsterland dashboard overview (dark theme) Lobsterland dashboard overview (light theme)

What your crew handles

Built for the real busywork.

Talk to it anywhere

Message it like a coworker, from anywhere.

Give your agent a name and chat from the built-in dashboard, Slack, WhatsApp, or Telegram. It works while you're away and reports back when it's done.

See messaging setup →
app.lobsterland.ing/app/messaging
Telegram integration in Lobsterland (dark theme) Telegram integration in Lobsterland (light theme)

A crew, not one bot

Run a team of named specialists.

Spin up separate agents — a researcher, a QA tester, a BD rep — each with its own workspace and memory. Route each channel to the right one.

Multi-agent routing →
app.lobsterland.ing/app/agents
Multi-agent routing in Lobsterland (dark theme) Multi-agent routing in Lobsterland (light theme)

It drives a real browser

A browser your agent uses — and you can take over.

A persistent hosted browser keeps cookies and login state. The agent works through local CDP; you open the same browser from the dashboard for logins, MFA, or a manual nudge.

Hosted browser →
app.lobsterland.ing/app/addons
Hosted browser in Lobsterland (dark theme) Hosted browser in Lobsterland (light theme)

Set it and forget it

Schedule the busywork on cron.

Create, edit, run-now, and delete scheduled jobs from the dashboard — daily briefings, recurring scrapes, weekly reports. No crontab, no SSH.

Cron management →
app.lobsterland.ing/app/instance/cron
Cron job management in Lobsterland (dark theme) Cron job management in Lobsterland (light theme)

See every dollar

Per-model cost and live logs, in plain sight.

Token usage and spend broken down by model and time, plus live instance logs. You bring your own keys — no markup on model costs.

Usage & cost →
app.lobsterland.ing/app/instance/usage
Per-model usage and cost in Lobsterland (dark theme) Per-model usage and cost in Lobsterland (light theme)

Managed Paperclip

Paperclip, auto-attached to every instance.

Enable a managed Paperclip workspace once — it auto-attaches to every OpenClaw you have now and every one you launch later. No API keys, no endpoint wiring. $4.90/mo.

Managed Paperclip →
app.lobsterland.ing/app/paperclip
Managed Paperclip workspace in Lobsterland (dark theme) Managed Paperclip workspace in Lobsterland (light theme)

And everything else your crew handles.

Control plane

  • Workspace editor — edit files in the browser
  • Skills manager — install and toggle agent skills
  • Cron jobs — schedule recurring tasks
  • Environment variables — managed and encrypted
  • 500+ models — bring your own provider keys
  • Package managers — npm, pip and more, built in

Channels & access

  • Built-in chat — talk to it in the dashboard
  • Telegram, Slack, WhatsApp — optional channels
  • File uploads — send it docs and images
  • Hosted browser — it drives, you take over
  • Managed Paperclip — auto-attached workspace
  • CLI auth — sign in with your provider CLI

Security & ops

  • No public IP — isolated runtime by default
  • Encrypted credentials — secrets stay sealed
  • Allowlist access — only people you approve
  • Usage analytics — per-model cost and tokens
  • Export anytime — your data is yours
  • Addons — extend with optional integrations

Agent Kits

Skip the blank slate. Hire a specialist.

Each kit launches an OpenClaw agent with the workspace files, a managed toolbelt, and a first useful workflow already in place. Pick the role and put it to work.

Personal productivity · light tier

Personal Assistant

Turn email, calendar context, notes, and loose todos into a daily plan.

View kit → Product work · medium tier

Product Manager

Convert research, feedback, and product context into PRDs, roadmaps, and buildable tickets.

View kit → Operations · light tier

Project Manager

Keep projects moving with status reports, risk tracking, meeting notes, and follow-up drafts.

View kit → Research · medium tier

Research Analyst

Produce cited research briefs, competitor scans, source tables, and watchlist updates.

View kit → Finance research · medium tier

Financial Analyst

Structure financial research with filings, market data, reproducible analysis, and sourced notes.

View kit →
Browse all agent kits →

Security by architecture

Private by default, powerful inside.

  • No public IP. Your instance isn't reachable from the open internet — random scanning can't find it.
  • Allowlist-only access. Only the Telegram users, Slack channels, or WhatsApp contacts you approve can message it.
  • Encrypted credentials. AES-256-GCM at rest. Full root capability stays inside your isolated runtime.
  • No lock-in. Export your whole workspace as a .tar.gz anytime — your data leaves with you.

Security deep-dive →  ·  Compare with self-hosting →

app.lobsterland.ing/app/instance/messaging
Allowlist access control in Lobsterland (dark theme) Allowlist access control in Lobsterland (light theme)

From people running it

Rated 5/5 on F6S & G2.

“The easiest and fastest way to set up OpenClaw without running your own device at home 24/7. Super easy to use and very affordable.”
JeroenF6S
“It is a revelation! I was up and running with my personal instance in just a few easy clicks. Game changer.”
JeffF6S
“I was able to set up every integration I wanted. Very affordable, and I don't have to do any setup. It just works.”
Jeroen C., Co-FounderG2

Pricing

Lobsterland runtime capacity for managed agents.

No bundle math. Choose how many Lobsterland instances you need and pick the capacity tier for each runtime. Available surfaces depend on whether you launch OpenClaw or Hermes. LLM usage stays BYOK: you pay your model provider directly.

monthly bill = instance quantities × selected tiers + optional add-ons
Start here per instance

Light

Entry runtime capacity for personal assistants, browser chat, and light automations.

$6.90 /month
1 CPU
3 GB RAM
20 GB storage

Everyday assistant work and light channel activity

Start Light for $6.90/mo
Most flexible per instance

Medium

More headroom for larger runtime state, browser work, and busier agents.

$14.90 /month
2 CPUs
8 GB RAM
80 GB storage

Research, memory-heavy agents, and busier workflows

Start Medium for $14.90/mo
Heavy work per instance

Large

High-capacity runtime for demanding automations, larger state, and long-running agent work.

$29.90 /month
4 CPUs
16 GB RAM
160 GB storage

Heavier background tasks and larger operational envelopes

Start Large for $29.90/mo
Live estimate

Build a monthly runtime bill

Mix instance sizes, add Paperclip for OpenClaw workspace coordination, and see the hosting total before model-provider usage.

$6.90
$0.00
$0.00
Estimated monthly total $6.90 /month before LLM usage
Optional add-on

Managed Paperclip workspace

Provision a Paperclip workspace for OpenClaw agents with no API-key wiring, endpoint setup, or manual re-sync. New Paperclip deployments are $4.90/month; already deployed Paperclip workspaces are not interrupted.

$4.90 /month
One-time setup help

Need help getting one OpenClaw workspace live?

The managed setup experiment is intentionally narrow so it does not imply unlimited support: one user, one instance, one primary channel or add-on, one BYOK model path, and a short follow-up window.

Managed Setup Launch $199

One-time setup help for one user, one hosted OpenClaw workspace, one supported channel/setup path, first-run configuration, a smoke test, and a handoff note.

Includes one 30-minute onboarding/debug session or equivalent async help plus 7-day setup-specific follow-up Start Managed Setup Launch

Not included: custom workflow/code work, unlimited support, self-hosted server repair, emergency response, compliance commitments, or third-party provider outages.

Runtime surfaces vary by launch choice

Light, Medium, and Large are capacity tiers. OpenClaw instances can use the full operating surface: dashboard chat, Telegram/Slack/WhatsApp, workspace editor, skills, cron jobs, environment management, usage analytics, encrypted credentials, no public IP, and export anytime.

Hermes instances focus on browser chat, Telegram, Slack, and logs — on the same account, dashboard, billing, and launch platform. The heavier surfaces (WhatsApp, workspace editing, skills, cron, Hosted Browser, Paperclip, and multi-agent management) live on OpenClaw. Compare runtimes →

Start Light for $6.90/mo

Need custom capacity or a bigger operational envelope? Contact us. Business accounts can share purchased capacity across invited teammates.

FAQ

Clear answers.

Is my assistant exposed to the public internet?
No. Your instance has no public IP and is not directly reachable from the public internet. Traffic flows through a private network boundary — random internet scanning cannot reach your assistant.
Can anyone message my bot?
No. You configure an allowlist of approved Telegram users, groups, Slack users and channels, or WhatsApp contacts. Only people you explicitly approve can message the assistant.
Does isolation limit what my assistant can do?
No. Isolation protects network and runtime boundaries, but your assistant has full root-level capability inside its own environment. It can run code, install packages with pip/npm/brew, use the browser, and execute any workflow you define.
How fast can I start?
With your LLM provider credentials ready, the full setup takes about 2 minutes. You can chat in built-in chat immediately, then optionally connect Telegram, Slack, or WhatsApp step by step.
Is built-in chat available, or do I need Telegram/Slack/WhatsApp?
Built-in chat is available by default. You can use your assistant directly in Lobsterland without connecting any external messaging platform. Telegram, Slack, and WhatsApp are optional if you want external channels.
Do you offer uptime guarantees?
We operate with a 99.9% uptime target for the managed hosting layer. Actual availability can still be affected by third-party dependencies (for example Telegram, Slack, WhatsApp, or your chosen LLM provider) and planned maintenance windows.
How do I control token spend?
The dashboard shows token usage and cost broken down by model and time period. You can spot spikes immediately and tune your assistant's behavior accordingly. You supply your own API keys — we don't add markup on LLM costs.
Can a team share one billing account?
Yes. Business accounts let one owner invite members by email, keep billing on the owner's account, and share the same purchased instance capacity across the team. Members still deploy from their own accounts. See business accounts →
Does Lobsterland include a hosted browser?
Yes. In Addons → Hosted Browser, enable a persistent browser sidecar for the instance. The agent uses it through local CDP, cookies and login state can persist on the instance volume, and you can open the same browser from the dashboard for auth, MFA, CAPTCHA, downloads, or manual takeover. See OpenClaw with browser →
Is Chrome Extension supported?
In Addons → Chrome Extension, generate a short-lived relay session, run the local node host command on the machine where Chrome is running, then authenticate extension options with OPENCLAW_GATEWAY_TOKEN. Finally, open the target tab and click the extension icon until badge shows ON to attach the tab for browser tasks.
Why not just self-host?
Self-hosting is valid for infrastructure-heavy teams who want full ownership. But it requires security hardening, maintenance, and operational burden. Lobsterland removes that work with private network defaults, isolated runtime boundaries, managed updates, and clear cost visibility. Read the full comparison →
Is OpenClaw the same as ClawdBot and Moltbot?
Yes — same core project, renamed over time. ClawdBot became Moltbot, and later OpenClaw. The technology lineage is continuous; the branding changed as the project evolved.
Does Lobsterland offer managed Paperclip hosting?
Yes. Lobsterland can provision a managed Paperclip workspace for your account — no external Paperclip account, no API keys, no endpoint wiring. Paperclip is an optional $4.90/month add-on for new workspace deployments. Enable it once from the Paperclip tab in your dashboard, and the workspace auto-attaches to every OpenClaw instance you have today and every one you launch in the future. New instances are mirrored automatically the moment they boot. Managed Paperclip hosting → · Feature page →

Move to managed OpenClaw
without starting over.

Import your current OpenClaw instance, then run it with secure defaults, clear usage tracking, and less ops overhead.

Import your current OpenClaw instance →
Cookie preferences