Legal
Acceptable-use policy
Approved customers must use the Teralor Tech ultra-fast, privacy-first inference API lawfully, safely, and within the service's technical limits.
1. Customer responsibility
Customers are responsible for their prompts, integrations, use of model output, and activity under their API key. Use must comply with applicable law, the terms of service, this policy, and rights held by others.
2. Prohibited use
Do not use the service to:
- break the law, facilitate fraud, or evade legal restrictions;
- infringe privacy, confidentiality, intellectual-property, or other rights;
- create, deliver, or meaningfully assist malware, credential theft, phishing, unauthorized access, or disruption;
- harass, threaten, exploit, or facilitate abuse of a person;
- generate or distribute unlawful sexual content or sexual content involving minors;
- circumvent authentication, model restrictions, prepaid controls, request limits, account blocks, or key revocation;
- probe or attack the service or connected infrastructure without prior written approval through the direct channel;
- resell, share, or transfer the assigned key or access without Teralor Tech's direct approval; or
- use output as the sole basis for medical, legal, financial, employment, housing, insurance, critical-infrastructure, or other high-consequence decisions.
3. Technical care
Keep requests within the published model, token, body-size, key, concurrency, and credit limits. Clients must not intentionally create abusive load, bypass denials, or repeatedly replay uncertain requests in a way that risks duplicate work or spend. Protect the bearer key and report suspected compromise without transmitting the key.
4. Enforcement
Teralor Tech may deny a request, revoke a key, block an account, or end access when it reasonably believes use violates this policy, threatens customers or infrastructure, creates legal or security risk, or exceeds approved bounds. Action may be based on non-content operational signals, customer reports, or credible evidence. The alpha does not promise continuous inspection of prompt or response content.
Questions about a proposed use or a block can be raised through the same direct channel used for approved onboarding. Access remains limited and review does not guarantee approval of a use case.