For creators

Your face and voice are your IP.Don't let deepfakes use them without paying you.

Signet monitors the open web for AI-generated impersonations of your likeness, drafts DMCA takedowns automatically, and gives you a cryptographic trail of every enforcement action.

What you're actually up against

Cloned voice in ads you would never approve
AI voice synthesis is good enough that 11 seconds of audio from one of your videos can produce a convincing ad for a sponsor you have never heard of. The first you know about it is when fans flag it.
AI-generated NSFW with your face
Deepfake adult content with your likeness is the #1 abuse vector creators report. Most platforms will remove it on a DMCA notice — if you can find every copy.
Deepfake "endorsements" of products that hurt your brand
Scammers paste your face onto sales videos for supplements, crypto, and "courses". Subscribers blame you for the refund-policy that does not exist.
Fan-cam edits scraping into the next training dataset
High-quality footage of your face and voice gets vacuumed into open-source training sets without consent. The next model knows how to be you.
Impersonator accounts running sponsorship-fraud or phishing your fans
A clone account with your face, a near-identical handle, and "official" in the bio reaches out to brands posing as you, or DMs your fans for crypto/giveaway scams. Meta's automated impersonation flow often sides with whichever account has more history. By the time you get a hearing, the damage to your brand and your fans is done.

How Signet protects you

Four layers, all included on Shield Pro. Each feature maps to one of the threats above.

Voice fingerprint monitoring
Read a short script during onboarding. We extract a voice embedding (the bytes are encrypted at rest, you can revoke any time) and look for matches across video and audio platforms.
Reverse face search
Continuous AI-image search across the open web — Reddit, TikTok, Twitter, Discord. Hits flag suspected face matches and deepfake-style edits separately.
Weekly automatic scans
Shield Pro scans every 7 days. Guardian scans daily. Both fire email alerts the moment a new detection lands so you can authorize a takedown before it goes viral.
Biometric-verified takedown filings
One click + mouse-signature, and we file on your behalf — DMCA for copyright cases (stolen footage, voice clones used commercially), platform Trust-and-Safety reports for impersonator accounts. Every filing carries your government-ID-verified identity record + biometric match. That verification chain is what platforms respond to differently — third-party institutional proof, not just a victim's own claim.

In practice

YouTuber, 500k subs
Weekly scans surface 3–5 deepfake re-uploads per month. Most are on third-tier video sites that respond fast to DMCA. Net cost: ~10 minutes per week reviewing and authorizing.
Voice actor
Voice fingerprint flags a TikTok ad for a sleep supplement using a clone of their voice. Detection-to-takedown: 14 hours.
Podcast host
Opt-out submissions to OpenAI, Anthropic, Stability, and Adobe stop new training runs from including their voice. Existing models cannot be untrained, but the bleed is plugged.

From early users

Signet is in private beta. Real testimonials replace these placeholders at GA.

“Quote from a happy beta creator goes here.”
— Beta user #1
“Quote from a happy beta creator goes here.”
— Beta user #2
“Quote from a happy beta creator goes here.”
— Beta user #3
Recommended for creators
Shield Pro
$29 / month
Weekly scans, voice protection, unlimited takedowns.
  • Weekly automatic scans
  • Unlimited takedowns
  • Voice fingerprint monitoring
  • AI training opt-out submission
  • Data broker removal checklist (top 20 US people-search sites)

Outgrowing Pro? Guardian ($99/mo) adds daily scans, a 24-hour support SLA, and an annual identity assurance report — useful if you have a press team or work with brands that require proof of likeness control.

Frequently asked

Do I have to give you my face and voice?

Yes — that is the thing you are protecting. The bytes are encrypted at rest with AES-256-GCM, the encryption key is versioned (we can rotate without losing access to historical records), and you can delete the entire account at any time with a 30-day cool-off window before permanent deletion.

What if my voice changes over time?

Re-record any time. New voice fingerprints replace old ones in monitoring; previous records stay in your audit log so you can prove what was on file when.

Can I revoke biometric consent later?

Yes — GDPR Article 17 + CCPA right to delete. Revoking consent stops monitoring immediately and triggers deletion of the biometric data within 7 days. Your previously-filed takedowns are unaffected.

Do I keep working with my agent or label?

The account is yours. V1 is single-user; V2 will add the ability to delegate authorization (so your manager can authorize takedowns on your behalf). For now, share the dashboard if needed.

How is this different from Pimeyes or Stable Attribution?

Pimeyes is reverse search but no takedown automation. Stable Attribution is training-set tracing but no proactive monitoring. We do both, plus voice (which neither covers), plus we file takedowns on your behalf — and the filings carry our biometric identity verification, which DIY-from-victim filings cannot match.

What if Instagram or TikTok ignores my impersonator-account report?

Their automated systems weight account history heavily and frequently side with an impersonator who has been operating longer than your reporting account. Signet's identity verification (government ID + biometric match + audit-logged record) is third-party institutional proof. We attach it to every filing, which gets reports past the automated layer to human review more often than a victim's own report does. We don't override platform decisions, but the rate at which they actually look goes up materially.

Stop being the unpaid training data.

Identity verification + first scan in under five minutes. Cancel any time.

For creators — Signet · Signet