Privacy Policy
This policy describes how LLM Tech ("we") handles data when you use our inference API and website. LLM Tech is operated by a registered sole proprietorship (jednoosobowa działalność gospodarcza) in Poland and is subject to EU law, including the GDPR.
1. Prompt and completion content
- Zero content retention. Prompts, completions, and any content submitted to the inference API are processed in memory for the duration of the request and are not written to persistent storage.
- No training. We do not use your inputs or outputs to train, fine-tune, or evaluate any model, and we do not provide them to third parties for that purpose.
- No human review. Request content is not read by humans.
2. Operational metadata
To operate and protect the service, our infrastructure records limited technical metadata: timestamps, request sizes in tokens, response latency, HTTP status codes, and the IP address of the connecting client (in web-server and rate-limiting logs). This metadata contains no prompt or completion content and is retained on two schedules: billing records (timestamps, token counts, applied prices, request identifiers, and the IP address of the connecting client — typically a platform's gateway rather than an end user) are kept for up to 13 months as billing evidence and then deleted; web-server and rate-limiting logs are kept for up to 30 days for security purposes.
3. Website
This website is static. It does not set cookies, does not include analytics or advertising scripts, and does not use tracking of any kind. Standard web-server access logs (IP address, user agent, requested path) are kept for up to 30 days for security purposes. All assets, including fonts, are served from our own servers — the website makes no requests to third parties.
4. Legal basis (GDPR)
Where the GDPR applies, we process operational metadata on the basis of legitimate interest (Art. 6(1)(f) GDPR) in operating, securing, and billing the service. Prompt and completion content is processed solely to deliver the response you requested (Art. 6(1)(b) GDPR) and is not retained.
5. Sub-processors
Inference runs on GPU infrastructure leased from cloud providers located in the EU. Traffic between our edge and GPU nodes is restricted to what is needed to serve the request. Current sub-processors: Hetzner Online GmbH (Germany — edge server) and Verda (Finland — GPU infrastructure); neither has access to request content. The maintained list with roles is published at llmtech.eu/security.
6. Your rights
Under the GDPR you may request access to, rectification of, or erasure of personal data we hold about you (in practice: operational metadata associated with your IP address or account), and you may lodge a complaint with a supervisory authority. Because request content is not stored, it cannot be retrieved or exported after the request completes.
7. Changes
We may update this policy as the service evolves. Material changes will be reflected by the date above.
8. Contact
Data controller: LLM Tech (sole proprietorship, Poland). Contact: artem@llmtech.eu.