Last updated May 12, 2026.
This is the privacy policy for gpus4all.dev. It is written to be read, not to be decoded. If a sentence is unclear, email us and we will explain it and fix the sentence.
When you create an account, we store the email address you give us and the name of the company or project if you provide one. When you use the service, we store the data needed to operate it: your instances, your deployments, your API keys (as hashes), and your billing records.
We do not sell personal data. We do not use your prompts or model inputs to train models. We do not read your instance contents, your uploaded model weights, or your request payloads except to troubleshoot a specific problem you reported. We do not run advertising.
Requests sent to the inference API pass through our systems to serve your deployment. We retain logs of API calls for 30 days for billing and abuse detection, and the logs contain request metadata and token counts but not full request bodies. If you need logs retained longer or not at all, tell us and we will set that per account.
Account and billing data is stored in the EU. Instance data lives in the region where you create the instance: Frankfurt, Ashburn, or Singapore. Cross-region transfers only happen when you move an instance yourself.
You can request a copy of your account data, ask us to delete it, or export it at any time. We honor deletion requests within 30 days. To exercise any of these, email contact@gpus4all.dev. We respond to every request; we do not have a data subject portal because we do not have the headcount for a portal.
Access to your account is protected by your credentials. We use encryption in transit everywhere. We review access to production systems quarterly and on any suspected incident. If you find a problem with our security, email security@gpus4all.dev and we will fix it and tell you what changed.
If this policy changes in a way that affects you, we email you before it takes effect and keep the date at the top of this page accurate.