( AI LABS )

Train on likeness data, the right way.

The people inside your training data are real. Royall gives AI labs a clean supply of identity data with verified consent, full provenance, and licensing that pays the people behind it — so you can build powerful models on a foundation that holds up.

Why it matters

Train on data you're allowed to use.

The identities inside training data belong to real people. Royall turns that reality into infrastructure — a supply of likeness rights with verified consent, clear provenance, and compensation that reaches the people who granted it.

Consent you can prove

Every identity in your dataset carries a verified, timestamped record of the permission behind it — not a scraped file with no paper trail.

Provenance end to end

Know where each likeness came from, who granted rights, and under what terms. Lineage travels with the data through training and beyond.

Rights that hold up

Scope, territory, and duration are explicit and machine-readable, so your legal and safety teams can audit exactly what you're allowed to train on.

Value flows back

The people whose faces, voices, and names shape your models get paid. Ethical sourcing and fair compensation, built into the pipeline.

The shift

From scraped to sourced.

The old way

  • Scraped from the open web with no permission
  • No record of who's in the data or whether they agreed
  • Rights are a guess — exposure you can't quantify
  • The people behind the data get nothing
  • Takedowns and disputes hit after training is done

The Royall way

  • Sourced from people who opted in on clear terms
  • Every identity carries a verifiable consent record
  • Scope, territory, and duration are explicit and auditable
  • Revenue flows back to the people who granted rights
  • Revocation is handled upstream, before it becomes a problem

How it works

Consented data, from source to model.

Source

Consented data,
at the source.

Access a growing registry of people who've opted their likeness in for training, with the rights and terms they've approved.

labs.royall.ai
Royall order builder

License

License at
the scope you need.

Acquire training rights at the scope you need — a single voice, a cohort, or a whole category — with machine-readable terms attached.

labs.royall.ai/license
License order builder

Verify

Verified consent,
per identity.

Each identity resolves to a verified, timestamped consent record — confirm exactly what you're licensed to train on before ingestion.

labs.royall.ai/verify
Verified consent per identity

Attribute

Provenance,
end to end.

Provenance stays bound to the data through your pipeline, so any output can be traced back to consented, licensed sources.

labs.royall.ai/lineage
Provenance dashboard

Compensate

Value flows
back to the people.

Licensing revenue routes back to the people and teams behind each identity — automatically, with a clean audit trail.

labs.royall.ai/payouts
Royall payouts

What you get

Defensible training, by default.

Regulators, partners, and enterprise customers increasingly ask AI labs to show their data was sourced with permission. Royall makes that answer a record you can hand over — not a story you have to tell.

  • A verifiable consent record behind every identity
  • Exportable provenance and licensing audit trail
  • Fail-closed revocation handled upstream of training
  • Compensation routed to the people behind the data

Questions from labs

The things research teams ask.

Build models people can stand behind.

Get access to consented, licensed likeness data and give the people who make your models possible a real stake in them.