Zinn® AI Start
$14.99/month
or $149.90 a year
- NVMe storage
- 25 GB
- Bandwidth
- 1 TB
AI hosting
Run AI features on your hosting without renting a GPU
Serverless inference on open models, billed by the thousand tokens against one balance you can see and cap. Bring your own provider key or spend Zinn® AI credits — either way it is the same meter, and your application sits on the same account.
Get startedPlans from $14.99 a month
Inference you can afford to leave running
A GPU you rent is billed whether or not anyone is using it, which is why most AI features never make it past a prototype. Serverless inference is billed for the tokens you actually spend, so a feature that gets used twice a day costs what two uses cost. The models are open ones — Llama, Mistral and their successors — served through our AI gateway, which caches and fails over between providers so a single upstream having a bad morning is not your outage.
One meter, whichever key you use
You can bring your own provider key, in which case you pay that provider directly and we never touch the bill. Or you can spend Zinn® AI credits bought from us. What you cannot do is end up paying twice, because both routes are metered in one place with one running total and one cap — the number you set is the number that stops the spending, regardless of which key served the request.
Included on every AI plan
- A monthly allowance of AI credits, and a spend cap you set yourself
- Bring your own provider key instead, at no extra cost
- SSH, Git and scheduled tasks, so the application calling the model lives here too
- Free SSL, daily backups and on-demand snapshots
Plans and pricing
Three tiers on our own fleet. Every one includes SSH, Git and scheduled tasks, and a monthly allowance of AI credits.
Zinn® AI Builder
$28.99/month
or $289.90 a year
- NVMe storage
- 100 GB
- Bandwidth
- 5 TB
Zinn® AI Scale
$99.99/month
or $999.90 a year
- NVMe storage
- 500 GB
- Bandwidth
- 25 TB
Powered by 100% renewable energy
Every server we run, on every product, is powered by renewable electricity. Not offset after the fact — sourced that way.
- 100% renewable electricity across all our hosting
- Data centres running at a PUE of 1.1 to 1.2
- Sites on our shared platform pass the Green Web Foundation's checks — a third party you can verify for yourself
Questions people ask before they buy
Do I need a GPU, or do you rent me one?
Neither. Inference runs on serverless open-model endpoints and is billed per 1,000 tokens, so there is no machine to reserve and nothing to pay for while it is idle. Dedicated GPU capacity is not something we sell today — we would rather tell you that than rent you one you do not need.
Can I use my own OpenAI or Anthropic key?
Yes. Add it under AI credentials and requests are made with your key, billed to you by that provider. Your Zinn® credit balance is untouched. You can switch between your key and our credits at any time without changing your code.
How do I know what it is costing me before the bill arrives?
Usage is metered per 1,000 tokens and the running total is on your dashboard as it happens, not at the end of the month. You set a cap; when it is reached, spending stops rather than continuing and surprising you.
Is my data used to train anybody's model?
No. Requests go through our gateway to the model provider and are not contributed to training. If you bring your own key, the terms you agreed with that provider apply to those requests.
What is not included yet?
One-click AI environments such as Jupyter, PyTorch and Ollama, and hosting for long-running agents, are not part of these plans. They need a machine you have root on rather than a shared hosting account, so they belong on Zinn® Private Cloud and we would rather leave them off this page than sell you something this plan cannot run.