Sub-processors
Scalix World engages a small number of third parties to deliver the platform. This page lists every one of them, what they process, and where. We keep it current and will post changes here before they take effect.
This list is incorporated into our Data Processing Agreement.
| Sub-processor | Purpose | Data processed | Location |
|---|---|---|---|
| Hetzner Online GmbH | All compute, storage and networking. Our servers are dedicated hardware we operate ourselves. Also the destination for encrypted off-site backups. | All customer data stored on or processed by the platform | Germany / Finland (EU) |
| Hostinger | Mailbox hosting and outbound email relay for correspondence and service email, pending activation of our own mail server. | Email content and contact details of people who write to us or receive email from us | EU |
| Stripe | Payment processing for customers billed by Energy FW Ltd. | Billing contact details and payment metadata. Card details go directly to Stripe and are never stored by us. | EU / US |
| Razorpay | Payment processing for customers billed by Scalix World Pvt Ltd. | Billing contact details and payment metadata. Card details go directly to Razorpay and are never stored by us. | India |
| Amazon Web Services | Secondary authoritative nameservers for scalix.world. Primary DNS runs on our own hardware; AWS carries replica copies of the zone only. | DNS zone records. No customer content or application data. | US / India |
| OpenRouter (OpenRouter Foundation / Paragon Labs, Inc.) | Model aggregation for the third-party catalog of Scalix AI: requests to cards exposed through the third-party pass-through catalog are routed to the selected model provider via OpenRouter. The default free tier runs on Scalix-operated infrastructure and is unaffected. | Prompts and inputs you submit to a third-party catalog model, and the generated output. Only for models you explicitly select from the third-party pass-through catalog. OpenRouter and the model provider process the request under their own published policies; their retention and training practices are theirs and vary by model. Scalix does not use this data to train models. | US |
AI inference
Scalix AI's default tier runs on Scalix-operated infrastructure. Third-party catalog models are pass-through: when you select one of those models, your request is routed through OpenRouter to the model provider, which processes your prompt and returns the response. OpenRouter is listed above as a sub-processor for that flow, and each model provider's own terms also apply to your use of its model. If you supply your own provider key, that provider processes your inputs under your own terms with it, not as our sub-processor. Training and data use: neither Scalix nor its sub-processors use your prompts, inputs, or outputs to train models; model providers in the third-party catalog are engaged by OpenRouter and their own training and retention policies apply to the model you select. Our inference usage records keep metadata only (model, token counts, cost), not prompt or output content. Separately, your infrastructure data (databases, object storage, functions) stays on our own hardware and is never sent to OpenRouter or any model provider; it reaches AI inference only if you submit it to Scalix AI yourself.
Not sub-processors
For clarity, because these are commonly assumed: object storage runs on MinIO, which is software we operate on our own hardware rather than a third-party service; and primary DNS is served by our own nameservers on our own hardware.
Changes
We will publish new or replacement sub-processors on this page before they begin processing customer data. Customers with a signed Data Processing Agreement may object under the terms of that agreement. For data protection enquiries or to be notified of changes, write to data@scalix.world.
Last updated: 2 September 2026