Everything about connecting, models, credits and privacy. Question not listed? Email us — you get an answer from someone who builds the platform.
Yes. The Router speaks the OpenAI protocol, and also Anthropic Messages and the OpenAI Responses API. You only switch the base URL and API key; your existing code or SDK keeps working without any refactoring.
Log in to your HostYourAI account and go to API keys. There you create a key; the base URL is shown right next to it: https://hostyourai.com/api/v1. The key is active immediately.
Through the same OpenAI-compatible endpoint: enter the Router base URL in your tool together with your API key. In n8n, pick an OpenAI node and replace the base URL there. Example configurations are in the documentation.
Use the OpenAI integration in those tools and replace the base URL with the HostYourAI Router URL, together with your own API key. For n8n, an example configuration is available in the documentation.
On the documentation page: the endpoints, examples per tool and the Sovereignty Mode settings.
Request the model hyai/auto and the Router picks the best model for your request itself. Handy if you do not want to pin a specific model; you can try it right away in the Playground.
The current list is in the model catalog and in your dashboard. The offering changes: models are added and sometimes phased out.
The Model Garden is updated continuously. Missing a model? Email info@hostyourai.com with which model you need and for what purpose — we will see whether we can make it available on the Router or as a dedicated deployment.
They vary per model and are listed with each model in the model catalog and the documentation.
That model is not live on the shared Router right now, but it can be made available on request, for example as a dedicated deployment. Email us which model and which use you have in mind.
Yes. All inference runs on GPUs in European data centers. On a default account, traffic can also be routed via Vast.ai and RunPod — US-incorporated GPU marketplaces that operate hardware on EU territory — for capacity and price reasons. If you want to exclude that entirely, enable Sovereignty Mode ("EU only") on your API key; then everything stays with fully European providers. Where every party is established and where it processes is listed on the sub-processors page.
The Router is shared capacity: you pay per token, with no fixed cost. A dedicated GPU is your own instance at a fixed hourly rate — interesting for constant load or specific model requirements.
You pay per token; prices per model are on the pricing page and with every model in the catalog. Top up a small amount, measure your real usage in the Playground or via the API, and multiply that by your expected volume.
In the dashboard under Billing. You can pay with the usual payment methods; your credits are available immediately after payment.
Yes. With auto top-up under Billing your balance is refilled as soon as it drops below your own threshold, so your credits never run out unexpectedly. You set the amount and the threshold yourself, and you can switch it off again at any time.
In the dashboard you see your balance, your usage per period and your transaction history. Individual requests are on the Activity page.
Partly. The Activity page shows which key made each request, but totals are aggregated per account (or per team). For a clean split per application, use a separate key per application.
Invoices are sent by email after each payment and are also available in the dashboard under Billing, including those for automatic top-ups.
Update your company details in your account settings; new invoices will then use those details. For corrections on an invoice that was already sent, email info@hostyourai.com.
No. Topped-up credits remain valid as long as your account exists.
Yes, with a team. You invite colleagues by email address and then share router keys and dedicated instances with the whole team. Usage is always billed to the account of the key owner; balance and invoices stay with that owner.
All inference runs on GPUs in European data centers. On a default account, traffic can be routed via Vast.ai and RunPod (US-incorporated GPU marketplaces, hardware on EU territory) alongside European providers, for capacity and price reasons. If you want to exclude that, enable Sovereignty Mode on your API key: then everything stays entirely with European providers. The full list is on the sub-processors page.
No. Prompts and outputs are not used to train models, and we do not log the content by default. Metadata such as model name, token counts and latency is kept for billing and abuse detection; which data is kept for how long is listed in the security overview.
Sovereignty Mode restricts your traffic to fully European infrastructure and European model providers. You enable it per API key with the "EU only" switch, also after the key exists. From then on the Router only routes via sovereign routes.
Yes, the standard DPA is available at hostyourai.com/legal/dpa. For a signed copy in your company name, email info@hostyourai.com with: legal company name, registered address, VAT ID or registration number, authorized signatory, DPO or privacy contact, and the email address of your account. We return it countersigned within two working days. For custom clauses we will get in touch.
Yes. The Playground in your dashboard lets you test any model directly in the browser, using your own credits, without connecting anything.
The dashboard is available in English and Dutch; you switch with the language picker in the top bar. The website is also available in German. Invoices and communication in English are available on request: let us know via info@hostyourai.com.
Email info@hostyourai.com from your account address. We delete your account and personal data; legally required retention periods, for example for invoices, still apply. If you still hold a balance, we refund the remainder to the payments it came from; manually granted credit expires.
A dedicated instance has no paused state: it exists or it does not. You delete it in the dashboard or with DELETE /api/v1/instances/{id}, and billing stops at that moment. Note: a running deployment keeps costing money, even without traffic. Every deployment stays visible in your History, with the minutes it ran and the amount actually charged.
If auto top-up is on, we first try to refill your balance. If that fails, your running dedicated instances are removed so the costs do not keep running, and the Router serves your requests again as soon as there is balance.
Email info@hostyourai.com or use the contact form. We usually reply the same working day.