1. Data covered by this policy
“AI data” includes prompts, scripts, audio, speech, voice samples, pronunciation rules, images, reference assets, model and voice selections, task parameters, generated outputs, ratings, corrections, safety signals and related technical records.
2. Processing required to provide the service
When you start a task, OpenSpeaker processes the selected inputs and settings to route, generate, transform, store and deliver the output. The minimum necessary data may be transmitted to the connected provider selected by the tool or routing logic.
This service processing is distinct from model training. A provider may have its own role and data terms; OpenSpeaker does not promise that every provider follows the same retention or training practice unless a current contract expressly guarantees it.
3. Evaluation, improvement and training
OpenSpeaker may analyze task success, latency, errors, aggregate usage and user feedback to secure and improve the platform. We may use content that is lawfully eligible—such as content separately contributed with valid permission or data that has been genuinely aggregated or de-identified—to evaluate, fine-tune or train OpenSpeaker models and quality systems.
Where identifiable content or personal data requires consent for training, the choice must be clear, separate and voluntary. Declining optional training use will not prevent ordinary generation. Consent may be withdrawn for future eligible processing through in-app support or vivoo.global.network@gmail.com.
OpenSpeaker keeps early-stage pricing low to learn from real usage and feedback. That commercial model does not remove the need for valid permission when personal data or protected content is used for training.
4. Voice cloning and synthetic identity
You may clone or transform a voice only when it is your own voice or you have valid, documented permission covering the intended use. OpenSpeaker may request proof, restrict a voice, remove material or suspend an account where consent, identity or safety is disputed.
Do not use synthetic voice, image or audio to impersonate, deceive, defraud, harass or manipulate. You are responsible for any disclosure, watermark, label or consent required when publishing AI-generated or synthetic media.
5. Data you must not submit
Unless OpenSpeaker expressly provides a suitable protected workflow, do not submit:
- Passwords, authentication secrets, payment-card data or private encryption keys.
- Government secrets, trade secrets or information subject to a confidentiality duty that does not permit this processing.
- Personal or biometric data collected without a lawful basis and required notice or consent.
- Copyrighted, licensed or private third-party material you are not authorized to process.
- Material intended for fraud, unlawful surveillance, deceptive impersonation or harm.
6. Safeguards and human review
We use access controls, logging, provider restrictions, abuse detection and task-level records to reduce misuse and improve reliability. Authorized staff or contractors may review limited content when necessary to investigate safety, support, fraud, quality or legal issues.
Where reasonably practicable, training and evaluation datasets are minimized, filtered and de-identified. De-identification is not treated as complete when a person can reasonably be re-identified from voice, face, metadata or combined records.
7. Rights, deletion and output responsibility
You retain rights you lawfully hold, subject to provider terms and applicable law. You grant the limited processing license described in the Terms of Service. AI output may not be unique and may require human review, attribution, disclosure or clearance before use.
You may request access, correction, deletion, restriction, objection or withdrawal of consent as provided by law. A withdrawal applies to future eligible processing and may not undo lawful aggregate results or training already completed where removal is technically impossible or not legally required.
8. Changes and contact
We will update this policy when training practices, connected providers or legal requirements materially change. Contact vivoo.global.network@gmail.com or use in-app support for consent withdrawal, data-rights requests or suspected misuse.
Legal and technical references