CryptoServers
Launch server

Dev sandbox / CI runners on offshore hardware

Self-hosted GitHub/GitLab runners, ephemeral preview envs, build farms — on hardware you didn't lease through a corporate billing system.

See plans from $54.99/mo Deploy now
No KYC, ever DMCA ignored No traffic logs Live in 60 seconds
1 Tbps
DDoS absorption
41s
Median deploy
3
Recommended locations
$54.99
Entry plan / month
Built for the workload

Why Dev sandbox / CI runners runs better here

Specific technical alignment, not generic copy. Each point below is something the workload needs and we provide by default.

Full KVM, nested virt, kernel modules. Docker-in-Docker, Kata Containers, Firecracker all work.
41-second deploy median. Spin up an ephemeral env per PR in less time than your test suite runs.
AMD EPYC 9454P on Scale tier — the same hardware that powers GitLab.com's shared runners.
No corporate AUP gotchas. Run anything from kernel fuzzing to chaos-engineering chaos.
Crypto-native billing — pay by the month, no payment-failure surprises in your CI pipeline.
Workload notes

What you should know about Dev sandbox / CI runners on Cryptoservers

Self-hosted CI runners (act, Earthly, Drone, GitLab Runner, GitHub Actions self-hosted) avoid the per-minute meter, the security boundary problems with cloud runners, and the slow restart times of GitHub-hosted-runner cold-starts. KVM with full root + nested virt + custom kernel modules means anything Docker-in-Docker, Kata, or microVM (Firecracker) runs without special handling. Ephemeral preview environments per pull request fit cleanly on vps-business with a Docker layer.

For a dev team running CI on cryptoservers, the unit economics are strong: a vps-pro at $54.99/month equates to roughly 110 GitHub-hosted runner-hours/month — but you get continuous availability (not metered minutes) and full hardware (not vCPU shares). Scale up for build farms by spinning multiple instances across our 4 jurisdictions and load-balancing at the runner registry.

Jurisdictions

Recommended locations for Dev sandbox / CI runners

Each location below has a specific legal or technical alignment with this workload — the others stay supported, but we suggest these first.

Netherlands

Amsterdam (NL-AMS)

Tier-1 carriers, lowest median EU latency, AMS-IX peering.

Dev sandbox / CI runners in Netherlands

Switzerland

Zürich (CH-ZRH)

Swiss court order required — foreign DMCA ignored by law.

Dev sandbox / CI runners in Switzerland

Romania

Bucharest (RO-BUH)

Constitutional court rejected EU data retention twice.

Dev sandbox / CI runners in Romania
Plans

Plans sized for Dev sandbox / CI runners

Pre-filtered by CPU, memory and bandwidth profile. Pick the cheapest that fits your projected load — you can hot-resize later without downtime.

VPS

Pro

Intel Xeon Gold 6430
  • 8 vCPU
  • 24 GB DDR5 ECC
  • 400 GB NVMe
  • 5 Gbps unmetered
$54.99/mo
Deploy
VPS

Scale

AMD EPYC 9454P
  • 12 vCPU
  • 32 GB DDR5 ECC
  • 640 GB NVMe
  • 10 Gbps unmetered
$69.00/mo
Deploy
Dedicated

Shield

AMD EPYC 7313P
  • 16 cores · 32 threads
  • 64 GB DDR4 ECC
  • 2× 1 TB NVMe
  • 1 Gbps unmetered
$79.00/mo
Deploy
Sysadmin FAQ

Dev sandbox / CI runners — questions answered

Will Docker-in-Docker work on KVM here?
Yes — we ship CONFIG_KVM and nested virt enabled on every host. dind, sysbox, gVisor, Kata Containers and Firecracker microVMs all work without sysctl edits or kernel rebuilds.
How many concurrent CI jobs can I run on a vps-pro?
Depends on the workload. Typical Go/Node/Python builds: 6–10 concurrent on vps-pro's 8 vCPU. C++/Rust LTO builds: 2–3. Browser-based E2E with headless Chromium: 4–6. Beyond that, Scale tier or ded-shield for sustained load.
Can I run GPU workloads or just CPU?
CPU only — we don't offer GPU instances. For ML CI or rendering, a different host is the right answer. Most CI doesn't need GPU.
How do I keep the CI secrets safe on a remote host?
Same way you would on any self-hosted runner: encrypt at rest (LUKS), use ephemeral runners (re-imaged per job), keep the long-lived secrets in your CI orchestrator and let the runner pull them at job start. We don't have visibility into the runner instance — full root means full responsibility.
What if I need to scale up for a release week?
Hot-resize works on every VPS plan — vCPU and RAM are live; disk needs one reboot. Or spin a temporary scale instance for a week and destroy it after. Our crypto billing doesn't penalize partial-month usage; you pay for what you use proportionally.
Other workloads

Run something else on the same hardware

Same network, same posture, same prices. Pick what you actually need to deploy.

Tor relay & exit hosting

Tor relay

Self-hosted VPN server

VPN server

Self-hosted mail server

Mail server

Bitcoin full node

Bitcoin node

Monero full node

Monero node

Lightning Network node

Lightning node

Matrix homeserver hosting

Matrix homeserver

Game server hosting

Game server

Offshore seedbox

Seedbox

Dev sandbox / CI runners — deploy in 60 seconds

No email, no ID, no account. Pick a plan, pay in crypto, get root.

Launch a server See the network