About Royall

Taylor Swift filed to trademark her voice this spring because AI clones were already good enough to fake it. SZA found 238 of her songs in training sets she never agreed to. YouTube opened likeness detection to every adult. The FTC started sending warning letters to "nudify" sites under Take It Down. New York made synthetic-performer ads disclose they're fake. The EU started requiring deepfake labels on August 2.

None of that is a niche creator issue. It's what happens when a face, a voice, and a name can be copied at scale, and there is still no simple way to say: this is me, this use is allowed, this one isn't.

That's why we built Royall.

Who we are

Anne Kavanagh, Jennifer Cullen and UJ Rosas have been working together, in different combinations, for years.

Jen was Anne's first boss more than two decades ago. They became colleagues, collaborators, and co-founders.

Anne met UJ in 2014 online. They made podcast artwork for a client. One freelance project turned into a working relationship that's lasted more than a decade.

We got into music by accident. Jen started managing her brother, Keith Cullen. My New York agency became his, and UJ and I came with her. We didn't have a radio budget. Then, at CES, UJ and I got in a rideshare. The driver was playing music, told us it was his friend's song, and sold us on the artist from the front seat. That became Steereo: pay drivers to introduce music instead of burning everything on radio.

We spent years after that inside music, entertainment, and audio. Tens of thousands of songs. Later UJ and I joined a podcast network and tech company, and for the last three-plus years we've been inside the plumbing of creator-led advertising and monetization, where hundreds of millions of people's work and income sit. We weren't looking to start an AI company. Generative models just got good enough that we couldn't tell anymore. Ear first, then eye.

The technology could recreate the person. The infrastructure for the person had not caught up.

What we noticed

Hollywood already has pieces of this. SAG-AFTRA wrote replica consent into a four-year deal. Estates are licensing voices (Stan Lee on ElevenLabs). Congress has NO FAKES on the Senate calendar, a federal right in voice and likeness that still isn't law.

The 1% can hire lawyers and file trademarks. Everyone else gets a deepfake and a takedown form.

That's the gap. Identity is an asset class now. It needs its own record, the way a song has an ISRC and a house has a deed. Music already had messy ownership: masters, publishing, splits, different entities. RAID, a Royall Asset ID, is that idea for a person.

This is not only a creator problem. If you have a face, a voice, or a name on the internet, you have the same exposure. A consumer app and an enterprise product will never reach that many humans on their own. Scale is an API any company can call: is this person claimed, is this use allowed, on what terms.

What we believe

People should choose how, when, and if they can be used.

Consent has to be cheap enough for a normal person, not a luxury for the 1%.

We shouldn't have a business that fights the thing we promise you: ownership and consent.

Royall is the registry, the monitor, and the live permission check. Register your face and voice. Set the terms. Get told when you show up somewhere you didn't agree to. Let a platform actually ask before it uses you.

Steereo started with a driver playing his friend's song. Royall started when we understood the person was becoming the content, and that this was everyone's problem.