Licensed sources only
Footage, imagery and voice used in training comes from rights-cleared, commercially licensed datasets and partner agreements - no unauthorized scraping of creators, brands, or public footage.
Trust & security
Built on licensed data. Guarded by design.
Every model, every asset and every export is traceable - so brand, legal and security teams can say yes to AI creative with confidence, not caveats.
Data sourcing
Licensed and synthetic data. Never scraped.
SuperMaya's models are trained on data we have the rights to use - not content pulled from the open web without consent. That distinction shapes every model decision we make.
Synthetic augmentation
Generated and simulated training data fills gaps without relying on any single brand's or creator's likeness, reducing exposure and bias versus scraped-only pipelines.
Your assets stay yours
Brand kits, product footage and prior campaigns you upload are used only to generate your ads. They are never used to train shared or third-party models.
Watermarking & traceability
Know exactly where every frame came from
Every export carries an invisible content credential and a visible-on-request AI disclosure, so provenance survives re-uploads, re-encodes and platform hops.
Generate
Signed at export
Every rendered video is stamped with a C2PA-style content credential recording model version, prompt lineage and timestamp.
Publish
Disclosure preserved
"AI Generated" tagging travels with the asset across platforms, so downstream viewers and regulators can verify origin.
Audit
Traced on request
Trace any published creative back to the brief, brand checkpoint version and generation run that produced it.
Enterprise controls
Governance that satisfies security review
Built for teams where legal, brand and IT all need a seat at the table before creative ships.
Role-based access
Granular permissions for creators, brand reviewers, media buyers and admins - separate who can generate, approve, publish and manage billing.
Full audit trails
Every generation, edit, approval and publish action is logged with user, timestamp and version - exportable for compliance review.
Brand guardrails, enforced
The Brand Vault blocks off-brand colors, claims and logo misuse before content ever reaches a reviewer's queue.
Enterprise contract terms
DPAs, custom data-retention windows, SSO/SAML, and dedicated environments available on Scale and Enterprise plans.
Private brand workspaces
Secure, isolated spaces per team or brand - assets, brand vaults and generations never cross workspace boundaries.
Review & approval
Built-in review cycles and sign-off before anything is published, with a full history of who approved what.
Certifications
Compliance, in progress and in place
We publish status honestly rather than implying certifications we don't yet hold.
GDPR-aligned data handling
Encryption in transit & at rest
SOC 2 Type IIIn progress
FAQ
Security questions, answered
Do you train models on my brand's assets?
No. Anything you upload - product footage, brand kits, past campaigns - is used only to generate your own ads. It is never added to shared or third-party training sets.
Can I tell which ads were AI-generated after the fact?
Yes. Every export carries an embedded content credential recording model version and generation lineage, independent of the visible "AI Generated" tag.
Do you support SSO and custom data retention?
SSO/SAML and configurable retention windows are available on Scale and Enterprise plans, along with a signed DPA.
Where can I request an audit log export?
Admins can export the full audit trail from account settings, or request a scoped export from your account team for a compliance review.