Its own machine
An always-on dedicated VPS. Not a shared container. Not your laptop. Not a tab you forgot to close.
A first principles approach to making agentic organizations.
$16/mo · Price locked for life|You don't pay until it's live
A real machine. Not a prompt box.
A dashboard, not SSH
See what your operator is doing, restart it, check its health, and steer it from a real UI.
Working templates
Pre-built operator setups so day one is shipping work, not configuring a blank file.
We do the setup
Provisioning, dependencies, networking — handled. You get a running operator. You don't pay until it's live.
Loved by founding operators
What operators are saying.
Founding operator beta. Quotes shown here are placeholders pending real attestations.
“First time my agent has actually survived me closing my laptop. Whole different feeling shipping work this way.”
Placeholder Beta User@founding_op_001
“I was paying for 6 different AI subscriptions and a $20 VPS I forgot existed. lol.gl just collapses all of that.”
Placeholder Beta User@founding_op_002
“Giving my operator its own phone number and inbox changed how I think about delegation. It's not a chatbot anymore — it's an employee.”
Placeholder Beta User@founding_op_003
“The dashboard alone is worth it. I finally know what my agent is doing instead of grepping logs over SSH at 11pm.”
Placeholder Beta User@founding_op_004
“Took me longer to fill out the waitlist form than to spin up a working operator once I got in.”
Placeholder Beta User@founding_op_005
“Pointed my Hermes install at lol.gl and it just worked. Persistent, always-on, with a real wallet.”
Placeholder Beta User@founding_op_006
Every AI, one seat
No more juggling 8 AI subscriptions.
No need to keep up anymore with who has the best models. Automatically use the best AI for the job across seven different modalities — text, code, image, video, music, voice, and biology.
No need for eight different AI subscriptions. Use every AI, now all in one place, on one bill.
- Text · Code · Image · Video · Music · Voice · Biology
- Auto-routed to the best model per task
- Bring your own API keys on Pro for zero-markup usage
Laptop
Desktop, shell + filesystem
Phone
SMS, calls, voicemail
Email
Send, receive, reply
Keys
Secrets, permissions & auth
Wallet
Credit card, budgeted spend
Context
Persistent memory
Agents that get work done
Give your AI the tools humans use.
Real tools, not just a prompt box. Your agent has a shell, a filesystem, a phone number, an inbox, secrets, and a spending budget — everything a human operator would need to actually ship work.
You're no longer the blocker for your business. Create an agent, give it a goal, and it gets real work done while you sleep.
Bring your own agent
Any agent. One machine.
lol.gl is unopinionated about which agent runs on it. Run OpenClaw, Hermes, Claude Code, Codex, Cursor background agents, Paperclip orgs, or your own — on a real box with a phone, a wallet, an inbox, and persistent state.
OpenClawContinuous agent
HermesSelf-improving
Claude CodeAnthropic
CodexOpenAI
CursorBackground agents
PaperclipOrg orchestration
Bash / HTTPAnything scriptable
Your ownMIT-licensed agents welcome
If it can run on Linux, it can run on lol.gl.
operator ~ /home/agent
$ uptime
14 days · load 0.21 0.17 0.12 · 6 tasks running
$ df -h /home/agent
Size 80G · Used 12G · Avail 68G · 16%
$ ip addr show eth0
inet 185.199.108.42/24 · you own this IP
$ sudo whoami
root
$ neighbors
none. 1 operator, 1 machine.
Your own dedicated machine
Not a shared sandbox. Not a chat tab. A real box that's yours.
Every subscription runs on its own dedicated VPS. Full root. Full filesystem. Full network. No noisy neighbors, no shared GPUs, no someone else's prompts leaking into your context.
Your operator's work stays yours. Encrypted at rest, backed up daily, and behind DDoS protection from day one.
- Dedicated VPS — 4–8 vCPU, 8–16 GB RAM, 80–160 GB SSD
- Full-disk encryption + daily snapshots
- Europe or US region (you choose)
Identity
A real machine for your AI operator. Nothing more, nothing less.
What lol.gl is
- A dedicated machine your AI operator owns
- A control surface to see, restart, and steer it
- The phone, wallet, inbox, and keys it needs to do real work
- A managed runtime — provisioning, networking, and snapshots handled
What lol.gl is not
- Not a chat tab you forgot to close
- Not a shared sandbox or playground
- Not a toy demo or generic AI builder
- Not an agent framework — bring your own (OpenClaw, Hermes, Claude Code, Codex)
Before & after
What changes when your operator has its own machine.
Without lol.gl
With lol.gl
Your agent dies the moment you close the tab or shut your laptop.
It runs on its own always-on box. Sleep, travel, reboot — it keeps working.
You have 20 chat tabs open and can't tell which agent is doing what.
One control surface. See every operator, what it's doing, and how to steer it.
Your agent's secrets and API keys live in your shell history and dotfiles.
Encrypted, scoped per-machine, rotatable from the dashboard.
Eight AI subscriptions, three API consoles, two phone numbers, no single bill.
Every model, a phone, an inbox, a wallet — one machine, one bill.
You're babysitting a chat window or hoping your $5 VPS doesn't fall over.
Managed provisioning, daily snapshots, and DDoS protection from day one.
One more thing · Coming soon
One operator.
An entire org.
Your machine becomes the office. Hire a CEO, a CTO, a researcher, a marketer. Each runs as its own agent. They delegate up and down the org chart. You approve the big calls. They ship the work.
YouThe board
CEO AgentStrategy · approvals
CTO AgentEngineering
CMO AgentMarketing
COO AgentOperations
EngineerIC
EngineerIC
WriterIC
DesignerIC
ResearcherIC
SupportIC
Hire roles, not tools
CEO, CTO, CMO, engineers, researchers — each with a job description and reporting line.
Goals cascade
Mission → project → agent → ticket. Every task knows what it's for.
Budgets per role
Hard-stops on spend. Soft warnings before the wall. No runaway loops.
You're the board
Approve hires. Approve strategy. Pause, override, or terminate anything.
Swarms ship after the dedicated machine, control surface, and tooling layer are stable. Founding operators get them first.
From signup to usable operator.
01
Join the beta
Sign up and claim your founding price. Small cohort, direct support.
02
We provision your machine
lol.gl spins up a dedicated VPS, installs the operator stack, and configures networking.
03
Track setup status
Watch provisioning progress in the control surface. No guessing. You only pay once it's live.
04
Start from a template
Pick a starter template, customize it, and your operator is running.
Plans
Pro private beta. Basic and Sovereign roadmap.
Private beta access centers on Pro — always-on agents with comms, a budgeted wallet, BYOK, and swarm capabilities on dedicated infrastructure. Basic and Sovereign tiers will follow as onboarding expands.
Basic
$0/mo
Pay-as-you-go after credits
Every model in one chat. No subscription.
- Unified chat across every major AI model
- Personal AI Agent
- 5 free credits upon signup
- Pay-as-you-go after free credits
Pro
TBD
Includes everything in Basic
Always-on agents, BYOK, comms, and swarms.
- Everything in Basic
- Agents get a Phone, SMS, and Email
- Agents get a Budgeted Wallet & Card
- Bring your own API keys (reduces cost)
- Run local models
- Agent swarm capabilities
- Agentic Organization Capabilities
Sovereign
Custom
Single-tenant runtime
Everything in Pro, on dedicated infrastructure you control.
- Everything in Pro
- Dedicated single-tenant runtime
- Region pinning and data residency
- Custom security and compliance review
- Dedicated support
Join the waitlist for Pro private beta invites and future Basic and Sovereign availability.
You either have your own machine.
Or you're borrowing someone else's.
FAQ
Frequently asked questions.
How is this different from running my agent on a $5 DigitalOcean VPS myself?
The VPS is the easy part. lol.gl is the surrounding system: a phone number, an inbox, a budgeted wallet, encrypted secrets, rotated backups, a control dashboard, and pre-tuned operator templates. You're not paying for compute — you're paying for the parts that take a weekend to wire up and a year to keep working.
Can I run OpenClaw, Hermes, Claude Code, or Codex on it?
Yes. lol.gl is unopinionated about the agent runtime. Bring OpenClaw, Hermes, Claude Code, Codex, Cursor background agents, Paperclip orgs, or your own scripts. If it runs on Linux, it runs on lol.gl.
What does “don’t pay until it works” actually mean?
You join the waitlist for free. We provision your machine and don't bill the monthly subscription until your operator is actually live and reachable on its own dedicated VPS. If we can't stand it up, you don't pay.
Is this a shared sandbox or a real machine?
A real machine. Your own VPS — full root, full filesystem, full network, no shared GPUs, no neighbors. Your operator's work, secrets, and state stay on a box that's only yours.
What happens to my data?
Encrypted at rest, daily snapshots, your choice of US or EU region. You can pull a full export of the machine state at any time. If you cancel, you take your data with you.
How does this relate to OpenClaw?
OpenClaw is one of the agent runtimes lol.gl supports out of the box — but it's not the only one. Think of lol.gl as the office, the keys, the wallet, and the phone line. Whatever agent you hire is the one doing the work.
Want early access?
Join the private beta waitlist. We'll send an invite when your workspace is ready.
Early access waitlist|Pro onboarding|Basic and Sovereign roadmap