Skip to content
[ Pricing ]

One license. Your infrastructure.

feder8d installs inside your own data centre or cloud — a flat monthly license, not per-seat. You run it on your own hardware; no data ever leaves your walls.

Dedicated (On-Prem)

$999/mo

Annual license · unlimited users · fully air-gapped.

  • Unlimited users & workspaces
  • Any open model — DeepSeek, Kimi, Llama, Qwen, Mistral
  • Runs fully air-gapped, zero egress
  • Role-based access, user groups & a full audit log
  • Your own GPUs, or we help you size them
  • Scales from a single box (Mac mini / NUC) to a full cluster
  • Signed bundle + Ed25519 license — no control-plane dependency

No card for the trial · pay online, get your installer by email · the license is the only thing that gates it.

Prefer a dedicated cluster we run for you? Dedicated (Managed) from $30,000/yr — talk to sales →

[ Frequently asked ]

How do I buy?

Online, self-serve. Fill in the purchase form, pay by card through our payment provider (Dodo Payments — card details never touch feder8d systems), and within minutes you receive an email with your installer download (license included). Extract, run ./install.sh, done.

How does the trial work?

Fill in the on-prem form and we issue a signed trial license inside a ready-to-run installer, on the spot. You run it on your own hardware for 14 days. Nothing is provisioned on our servers — the license is the only thing that gates the install.

Does it need to call home?

Air-gapped installs validate the Ed25519 license locally and never call out — they run entirely offline. A connected install re-checks the license periodically so month-to-month billing stays in sync; you can turn that off and run fully air-gapped instead.

What's in the license price?

One flat monthly fee for the whole company: unlimited users and workspaces, any open model, priority support, and signed bundles. No per-seat billing, no token metering. You provide the hardware.

Can I bring my own hardware / GPUs?

Yes — that's the point. Run it on your existing GPUs, or we'll help you size a box. If you'd rather we run a dedicated cluster for you, that's the Managed option.

Can I run it on a single box, like a Mac mini in the office?

Yes. The same license runs on one machine via a Docker Compose install — no Kubernetes — with the LLM on the host GPU (Apple Metal, or an NVIDIA box). Perfect for a small team that wants sovereign AI on a box in the corner. It scales up to a full cluster later with no change to your license. See the self-host guide.