Acceptable Use Policy

Unlawful or covert recording. Record, transcribe, monitor, or upload a communication without all notices, rights, and consents required by the laws applicable to every participant.

High-impact automated decisions. Use a score or Output as the sole or determinative basis for employment, compensation, termination, credit, housing, insurance, healthcare, education admission, legal services, or another legally significant decision.

Regulated or highly sensitive data. Submit protected health information, full payment-card data, consumer reports, government identifiers, passwords, private keys, precise geolocation, information about children, or special-category data unless expressly authorized in a signed agreement.

Spam or deception. Send unsolicited commercial messages in violation of law; use false headers, deceptive subjects, impersonation, fabricated testimonials, or undisclosed AI-generated communications where disclosure is required.

Discrimination or surveillance. Profile or treat people unlawfully based on protected traits, infer highly sensitive traits, conduct covert employee surveillance, or retaliate against a person for declining recording.

Security abuse. Probe, scan, disrupt, overload, bypass limits, introduce malware, access another tenant, extract credentials, or attempt to reverse engineer the Service except where law expressly permits.

Third-party rights. Infringe privacy, publicity, confidentiality, intellectual-property, contractual, or other rights; scrape or import data contrary to applicable terms or law.

Model extraction and resale. Use automated means to copy material portions of the Service, build a competing model or dataset from Outputs, resell access without written authorization, or remove ownership notices.

Dangerous or illegal activity. Facilitate fraud, harassment, threats, violence, exploitation, illegal goods, or evasion of sanctions or export controls.