Adaptive
Compute Platform
Compute that scales when you do.
European compute platform for modern workloads. Pay-per-Second billing on scale-to-zero infrastructure.
EU-hosted · Pay-per-Second · Scale-to-Zero
💰 Personal savings
4 GB RAM, 2 vCPU compute + storage
Our Mission
We're building an adaptive compute platform around one simple principle:
Compute should only cost when it's actively used.
The problem today
Our approach
Compute that thinks with you.
Deploy AI agents, game servers or microservices on scale-to-zero infrastructure.
Per-second billing, EU-hosted.
The Problem with traditional hosting
Why pay for compute that sits idle 80% of the time?
Adaptive compute.
| Feature | Hyperscalers / classic hosts Always-on billing | xape Pay-per-second, scale-to-zero |
|---|---|---|
💰Billing | 24/7 fixed cost or hourly | Per active second |
📊Idle cost | Full price at 0% load | €0 — workload pauses automatically |
⚡Deploy time | Minutes to hours | Under 60 seconds |
👥Scaling | Manual sizing, lock-in | Adaptive: scale-to-zero to cluster-grade |
🔋Workload running | 24/7 — even with no load | Only when actively used |
💳Data location | Often US / unclear | EU-hosted, GDPR-compliant |
🔍Transparency | Complex pricing calculators | Per-second, one price per workload |
🌱Efficiency | High idle costs | Up to 88% less idle consumption |
Always-on compute
xape
Done paying for idle?
Create an account and deploy your first workload with per-second billing.
Create accountHow easy it works
Three steps to your first workload
Create account
Sign up in under a minute — no credit card, no contract.
Deploy a workload
Deploy game servers, AI agents or microservices via dashboard or API.
Pay per second
Workload starts on demand, pauses on idle. You pay only the active seconds.
Self-service · No minimum term · EU-hosted
OpenClaw
AI assistant on adaptive compute
Telegram integration
Chat with your AI assistant on Telegram, anywhere, anytime.
Web dashboard
Full web interface to configure and manage your assistant.
Adaptive compute
Active on request, paused otherwise — you pay the active reply time plus OpenAI tokens.
Beta Prices after Launch
Three compute tiers, billed per active second. These plans apply to compute workloads.
Mini
Small services, hobby workloads
- 2 vCPU
- 4 GB RAM
- Scale-to-Zero
- Per-second billing
- Automatic backups
- EU-hosted
Small
Standard for microservices and game servers
- 4 vCPU
- 8 GB RAM
- Scale-to-Zero
- Per-second billing
- Automatic backups
- EU-hosted
- Self-service API
Medium
Production workloads, AI agents
- 8 vCPU
- 16 GB RAM
- Scale-to-Zero
- Per-second billing
- Automatic backups
- EU-hosted
- Self-service API
- Cluster-grade SLA
Persistent storage
Minimum: 10 GB (€1/month)
- NVMe storage, EU-hosted
- Automatic backups
- Per-second billing
Cost calculator (Small plan)
Per-second billing. No hidden fees, no setup cost.
OpenClaw — AI assistant hosting
Pay-per-second compute for your AI assistants
Basic
- Dedicated instance with web dashboard
- Telegram bot integration
- OpenAI GPT models
- 2 GB RAM, 10 GB storage
- TLS-encrypted, own subdomain
- Scale-to-zero between requests
+ OpenAI usage based on consumption
Service module
- Everything from Basic
- Dedicated point of contact
- Email automation
- Custom integrations
- Operational goal achievement
- Priority support
Frequently Asked Questions
What you should know about xape
Still have questions?
Self-service dashboard, Discord community or direct email support.