Generated Media & Voice Policy

1. Images, dictation, and voice

Zexir can offer image generation, microphone dictation, and voice conversations where available. The model, device, browser, permissions, and plan determine supported features. A failed request or unfinished stream does not mean an image or recording was successfully produced.

This policy applies to prompts, reference images, uploads, transcripts, generated media, and the way you use those results. The Privacy and Acceptable Use policies also apply. Review the final output before saving, sharing, or relying on it.

2. People and permission

Use another person’s image, voice, identity, or private material only with appropriate rights and consent. Do not create deceptive impersonations, non-consensual intimate material, exploitative depictions of children, fraudulent endorsements, or media intended to facilitate harassment.

Permission to possess an image or recording is not necessarily permission to alter, publish, commercialize, or train on it. Consider the subject, context, audience, and relevant laws. Restrictions apply even if the result is labeled synthetic.

3. Microphone and recording controls

Microphone access is controlled by browser and operating-system permission. Use it in a place where recording is appropriate and people nearby understand the situation. Dictation can capture background speech, names, and other sensitive information unintentionally.

Stopping a microphone session limits new capture through the product control; it does not necessarily delete audio or transcripts already processed. Revoke microphone permission in your browser or system settings when needed. Check visible recording indicators rather than assuming a connection is inactive.

4. Processing and storage

Media tasks may send prompts, uploaded references, images, audio, or transcripts to the providers needed to complete the request. Saved conversations, Library items, downloads, and provider records are different copies with different handling under the Privacy Policy.

Private mode does not make voice or image generation local. A local text model does not automatically imply local voice synthesis or image generation. Avoid recording passwords, payment details, confidential meetings, or regulated information without an appropriate supported arrangement.

5. Synthetic output and transcription errors

Generated media can depict events that never happened and can introduce inaccurate text, unrealistic objects, or misleading details. Speech recognition can mistake words, names, accents, or speaker identity. Voice responses can confidently repeat an incorrect answer.

Verify transcripts before sending them as instructions or publishing them. Do not treat generated media as evidence that a person said or did something. Use clear disclosure when omitting it would mislead an audience about authenticity or endorsement.

6. Publication and rights

Review rights to prompts, reference material, trademarks, music, likenesses, and generated output before publication. AI generation does not guarantee exclusivity, copyright protection, or freedom from third-party claims. The Intellectual Property & Takedown Policy describes reports about rights violations.

Do not remove required watermarks, provenance information, or license notices where law or the applicable provider’s terms require them. Do not present synthetic material as an official record, professional certification, or authentic communication from someone who did not authorize it.

7. Availability and safety controls

Generation may be declined, interrupted, or limited by safety checks, provider policies, technical constraints, or quotas. Repeated attempts to bypass a restriction can violate the Acceptable Use Policy. A safety control that permits a request is not legal clearance.

Latency and quality vary with provider and task. Some features need a network connection and continue consuming provider resources after an interrupted browser session. Consult the Billing Policy for usage and payment questions; do not assume a cancelled stream means no processing occurred.

8. Reporting harmful media

Report abusive or rights-infringing media to mail@zexiry.com with its safe location or request identifier and a description. Do not forward unlawful intimate or child-exploitation material to support; provide safe identifiers and use appropriate official reporting channels.

If media has already been published through an external service, report it there as well. We can review material within our control, but removing a Library item or chat does not erase copies held by others or undo a completed external post.

Related policies and contact

For related responsibilities, see our Acceptable Use, AI Accuracy & Reliance, Account & Security, and Subscription & Billing policies. Contact mail@zexiry.com for questions about this supplement.