Last updated: June 5, 2026
SceneFlow AI uses industry-leading cloud providers for generation, and applies its own layered guardrails on top—not as a pass-through service. Our approach combines Google-native safety on every generation path, Extended Creative Services with Guardrails when policy limits apply, optional Studio content validation, and forensic provenance for segment video.
This tiered model protects creators, reduces platform risk for our Merchant of Record operations, and maintains audit trails that support abuse investigations when required.
When creating a custom voice, the user must record themselves speaking a specific, randomized consent phrase. Speaker verification compares the consent recording to the uploaded voice sample.
If the voice in the consent recording does not match the uploaded sample, the clone request is rejected. Users cannot impersonate others without the person providing live consent.
Trust Statement:"Custom voice creation requires live consent validation and speaker verification before cloning is enabled."
Image and video generation runs through Google Vertex AI safety filters with thresholds tuned for professional creative workflows. Text generation uses configurable harm-category thresholds on the primary Google path.
When the primary Google path is blocked by policy after retries, SceneFlow may route to an alternate generation path. Content from that path undergoes mandatory pre-storage review before it is saved or delivered—subject to the same content standards.
Creators may run content validation across Blueprint, script, storyboard, and segment video using the same credit model as other Studio tools. Validation surfaces policy and risk signals as informational warnings—it does not replace your editorial judgment or export review.
High-risk content categories receive stricter enforcement for custom voices, including hate speech, fraud scripts, political disinformation, non-consensual impersonation, and content targeting minors.
Trust Statement:"Layered moderation—from Google-native safety through guarded fallback review and optional Studio validation—helps creators ship responsibly while reducing platform risk."
Mechanism: Access to curated stock voices only. Voice cloning is restricted for new accounts.
Guardrail: High-risk features are gated to reduce burner-account abuse.
Mechanism: Voice cloning unlocks after minimum account age, subscription eligibility, and trust-score requirements—including successful consent and speaker validation.
Guardrail: Account history and payment verification before enabling cloning capabilities.
Trust Statement:"High-risk features sit behind a trust wall. Voice cloning requires eligibility, consent validation, and speaker verification."
Successful segment video generations receive SceneFlow provenance records: a SHA-256 content hash, signed metadata (model source, user, project, timestamp), and optional in-file credentials embedding when enabled. Provider-native watermarks may also be present in generated media depending on the upstream model.
Provenance and audit logs support tracing content back to:
Trust Statement:"Segment video carries signed provenance metadata—supporting attribution investigations and platform accountability."
SceneFlow operates shared responsibility with its Merchant of Record (Whop). Google provides provider-scale safety on generation; SceneFlow adds platform guardrails, content validation, provenance logging, violation strikes, account suspension, and abuse reporting.
This section describes how SceneFlow enforces content policies. It is incorporated by reference into our Terms of Service (Section 10.1).
Moderation and enforcement events may record prompts, scripts, generation path, generic block-reason categories, and provenance hashes at delivery. These records reflect Service-delivered output only—they do not cover modifications made after export in third-party editors.
You may appeal a suspension by emailing abuse@sceneflowai.com within seven (7) days of notice. Reinstatement is not guaranteed. See our Terms of Service Section 12.2 for user liability and indemnification obligations.
Our Terms of Service explicitly transfer liability for generated content to the user. Users agree that they are solely responsible for:
We maintain audit logs of generation activity, including:
These logs may be disclosed to law enforcement upon valid legal request.
If you encounter content generated by SceneFlow AI that you believe violates our policies or is being used for malicious purposes, please report it immediately:
Consent & speaker validation: Custom voices require live consent recording and speaker verification.
Google-native safety: All generation runs through Google Vertex AI safety filters with production-tuned thresholds.
Extended Creative Services with Guardrails: Alternate generation paths require pre-storage review before delivery.
Additional moderation: Optional Studio content validation across project lifecycle stages.
Content provenance: Segment video receives signed hash metadata for forensic chain-of-custody.
Violation enforcement: Documented strike thresholds, suspension, and termination for repeated policy violations.
Liability & audit: Users remain responsible for generated content; SceneFlow maintains audit logs to support lawful requests.
Trust & Safety: trust@sceneflowai.com
Abuse Reports: abuse@sceneflowai.com
Legal Inquiries: legal@sceneflowai.com
Website: https://sceneflowai.studio
See also: Terms of Service | Privacy Policy | Refund Policy