Deploy Your Own
With One Telegram Command
Type /create in @OpenClawBoxBot. Pick a plan. Pay with Telegram Stars. Your secure, isolated OpenClaw instance deploys on Kubernetes in under 60 seconds.
❯ How It Works
From zero to your own AI assistant in four steps. No server management. No Docker. No nginx. Just Telegram.
Open the Bot
Find @OpenClawBoxBot in Telegram or click the link. No signup, no email, no account creation.
Type /create
One command. That's it. The bot shows you three plans — Starter, Pro, or Max — each with different AI models and resources.
Pay with Stars
Tap to pay with Telegram Stars. Native Telegram payment — no credit card, no Stripe, no redirect. Telegram takes 0% commission.
You're Live
Your isolated OpenClaw instance spins up on Kubernetes in under 60 seconds. Get your personal URL and start chatting with your AI across Telegram, WhatsApp, Discord, Slack, and more.
~/self-hosting-openclaw (4+ hours)
$ apt install docker nginx certbot
$ docker-compose up -d
$ certbot --nginx -d my.domain.com
$ ufw allow 443 && ufw enable
$ vim /etc/nginx/sites-available/...
$ systemctl enable --now openclaw
# configure DNS, SSL, updates, backups...
~/with-openclawbox (<60 seconds)
you: /create
bot: Pick a plan and pay with Stars
bot: ✓ Instance deployed at you.oclawbox.com
bot: ✓ Send me a message to chat with your AI
# done. that's it.
❯ Why OpenClaw Box?
The only managed AI hosting with one-command deploy, built-in model access, and Telegram-native payments.
1 Telegram Command Deploy
No server setup, no Docker, no nginx configs. Type /create in @OpenClawBoxBot, pick a plan, and your isolated AI instance is live on Kubernetes in under 60 seconds.
Pay with Stars & Crypto
No credit cards, no Stripe, no redirect pages. Pay with Telegram Stars right inside the chat — Telegram takes 0% commission. Crypto payments coming soon.
Cloud Hosted, 24/7 Available
Your instance runs on managed Kubernetes in the cloud. No downtime when your laptop sleeps. Always-on, auto-scaled, monitored, and backed up. We handle the infra so you don't have to.
GPT-5.2 API Included
We provide access to GPT-5.2, Grok-4, Claude, DeepSeek and 10+ models out of the box. No need to create OpenAI or Anthropic accounts, no API key setup. Just start chatting.
Where We Fit
We occupy the only position in the “low price + low friction” quadrant.
Official OpenClaw Team
We are OpenClaw. Same-day support for new releases, direct access to core developers, and auto-updates on day one.
MCP Servers in Every Plan
Connect Model Context Protocol servers on any subscription. Bring your own tools and data sources without upgrades.
Multi-Channel From Day One
Deploy once, chat everywhere — Telegram, WhatsApp, Discord, Slack, iMessage, Signal, Web, and Matrix.
❯ Architecture
Enterprise-grade infrastructure. Per-tenant isolation on Azure Kubernetes Service. Every instance gets its own namespace, network policy, and encrypted storage.
SQLite Database
Users, subscriptions, payments, tenants
Telegram Stars API
Invoices, pre-checkout, payments (XTR)
LiteLLM Proxy
genai.oclawbox.com — API key + budget per user
Azure Kubernetes Service (AKS)
tenant-alice/
OpenClaw Gateway Pod
Persistent Volume
NetworkPolicy (isolated)
tenant-bob/
OpenClaw Gateway Pod
Persistent Volume
NetworkPolicy (isolated)
tenant-carol/
OpenClaw Gateway Pod
Persistent Volume
NetworkPolicy (isolated)
Per-Tenant Isolation
Every instance runs in its own Kubernetes namespace with dedicated CPU, RAM, and storage limits. Network policies prevent cross-tenant traffic. Your data never touches another user's instance.
Automatic TLS & DNS
Traefik provisions Let's Encrypt certificates automatically. Cloudflare handles DNS. Every tenant gets a subdomain (you.oclawbox.com) with HTTPS on day one.
Auto-Updates & Backups
We roll out new OpenClaw releases with canary deployments — zero downtime. Daily backups protect your data. Rollback if anything breaks.
❯ System Design
A clear view of how requests move through OpenClaw Box and where your data lives. Built for isolation, auditability, and predictable spend.
Telegram
User message
OpenClaw Box
Routing + billing
Tenant Gateway
Your namespace
LiteLLM Proxy
Policy, quotas, provider routing
Model Providers
OpenAI, xAI, DeepSeek, BYOK
Tenant Storage
Logs, configs, memory, assets
Message Handling
Telegram messages flow to OpenClaw Box, then to your isolated tenant gateway. We only store routing metadata and subscription state in the control plane.
Model Requests
Requests are proxied through LiteLLM with per-tenant API keys and spend limits. BYOK traffic goes directly to your provider.
Tenant Isolation
Each tenant runs in a dedicated Kubernetes namespace with its own volume, network policy, and resource limits.
Data Retention
Chat logs stay inside your tenant volume. Control-plane data is minimal: user ID, plan, billing, and routing records.
❯ Simple, Transparent Pricing
Pay with Telegram Stars. No credit card. No Stripe. No hidden fees. Telegram takes 0% commission.
1 Telegram Star ≈ $0.013 USD. Buy Stars via Apple/Google IAP or @PremiumBot.
Starter
~$10/month
For developers and personal use
AI Models
Bring Your Own Keys — connect any AI provider (OpenAI, Anthropic, Google, Azure, local LLMs)
- ✓All messaging platforms
- ✓Your own subdomain (HTTPS)
- ✓Per-tenant Kubernetes isolation
- ✓Auto-updates & backups
- ✓Full OpenClaw feature set
Pro
~$25/month
For power users who want hosted models
AI Models
GPT-5.1, GPT-4.1, GPT-5-mini, GPT-oss-120b, Grok-4 Fast, DeepSeek V3.2
+ $25/mo AI credits
- ✓All Starter features
- ✓Hosted AI models via LiteLLM
- ✓Personal API key with budget tracking
- ✓/usage command to monitor spend
- ✓Priority email support
Max
~$100/month
For teams and heavy workloads
AI Models
All Pro models + GPT-5.2, GPT-5, Grok-4, Grok-4 Fast Reasoning, DeepSeek R1
+ $100/mo AI credits
- ✓All Pro features
- ✓Premium frontier models
- ✓Double the compute resources
- ✓Priority support
- ✓Extended storage (50 GB)
/create and enter the promo code.❯ Supported AI Models
Frontier models from OpenAI, xAI, and DeepSeek — hosted and managed for you via LiteLLM. Or bring your own API keys on the Starter plan.
OpenAI
xAI
DeepSeek
BYOK (Starter)
Pro and Max plans include hosted models via genai.oclawbox.com (LiteLLM proxy). Each user gets a personal API key with monthly budget tracking. Use /usage to check your spend.
❯ Telegram Commands
Everything is controlled via chat commands. No dashboard, no web portal — just Telegram. Here's the full command reference.
Getting Started
/createKey CommandCreate a new OpenClaw instance. Pick a plan, pay with Stars, and your AI deploys in 60 seconds. The only command you need to get started.
/startSame as /create. Also handles Telegram deep links for referrals.
/plansView all available plans (Starter, Pro, Max) with pricing, models, and resource details.
Manage Your Instance
/statusCheck your subscription plan, expiration date, instance URL, hosted models, and monthly credits.
/usageView your AI model usage and remaining credits (Pro & Max plans). See per-model spend breakdown.
/restartRestart your OpenClaw gateway. Useful if something gets stuck.
/logsFetch the last 20 lines of your instance's pod logs for debugging.
/upgradeUpgrade to a higher plan. Sends a new Stars invoice for the upgrade.
/downgradeDowngrade to a lower plan. Takes effect at the next billing cycle.
Account
/cancelCancel your subscription. Instance stays active until expiration, then gets suspended.
/deletePermanently delete your instance. Removes all data, revokes API access, ends subscription.
/referralGet your referral link. Earn bonus Stars when friends subscribe.
Support
/helpShow the full command list with descriptions.
/supportGet support contact info — Telegram channel, email, and payment support.
/termsView Terms of Service.
Direct Messaging
Beyond commands, just send any message to the bot — text, voice, photos, documents, videos, stickers, GIFs — and it gets forwarded directly to your OpenClaw instance. Chat with your AI naturally, right in Telegram.