Voice enrollment & your voiceprint
Everything starts with enrollment. You upload a few samples of your own voice, and our Resemblyzer speaker-embedding model (256-dimensional, GE2E architecture) converts them into a voiceprint — a mathematical representation of your vocal characteristics. This is the same class of model used by enterprise speaker-verification systems. We store the voiceprint, not a browsable copy of your recordings, and it is only ever compared against audio you ask us to check.
Your voiceprint is what makes the other three pieces possible: it is the reference an integrated platform checks a request against, the identity a provenance record attests to, and the baseline a verification compares a suspect clip to.
What VoiceSeal does not do: VoiceSeal does not crawl the web, does not continuously monitor platforms, and does not run automated daily scans for your voice. There is no web-wide clone search. We do not publish an accuracy percentage, because the honest answer depends on the audio and is confirmed by human review — not by a single automated number.
Pre-synthesis authorization
The most effective place to protect a voice is before audio is generated, not after it is published. A platform that has integrated VoiceSeal calls our pre-synthesis check API before it synthesizes audio with a given voice. The API returns your authorization decision and licensing terms for that voice, so the platform can proceed under license, or decline.
The check is designed to be unambiguous. Each response carries an explicit action — for example authorized, authorization_denied, allow_licensed, or no_registry_record when a voice isn't enrolled with VoiceSeal (in which case the platform makes its own decision). Enrolling your voice and setting default terms is what makes this check meaningful for platforms that adopt it.
The limit: pre-synthesis authorization only governs platforms that have chosen to integrate VoiceSeal. It cannot stop a platform that has not adopted the check. Growing that integration base is the point of our outreach mission below.
Provenance — trusted C2PA content credentials
When synthetic audio is marked through VoiceSeal, it carries an in-band, cryptographically signed C2PA manifest via the Trufo Provenance Platform. The signature is recognized on the C2PA trust list and is RFC 3161 timestamped, so any C2PA-aware tool can verify — independently of VoiceSeal — that a clip was marked and when. This is the standardized, machine-readable provenance mechanism referenced by the EU AI Act's Article 50 transparency requirements, which apply from August 2, 2026.
Acoustic watermarking is not a functioning control today. Our provenance guarantee comes from the trusted C2PA manifest, not from an inaudible watermark embedded in the waveform. You can review verifiable reference files and current status on our C2PA transparency page.
Rightsholder verification (Beta)
Verification is the after-the-fact tool, and it is deliberately you-initiated. You submit a suspect URL or file. We run Resemblyzer (GE2E) cosine similarity between that clip and your enrolled voiceprint, and every result is manually reviewed by our verification team before it is shared with you.
A similarity result — even a strong one — is a signal for your own review, not an automated verdict and not a legal determination that a use is unauthorized. A completed verification gives you a reviewed result and an evidence record you can use to pursue your own action; VoiceSeal does not file takedowns on your behalf.
The limit: verification only covers clips you choose to submit. A clone you never send us cannot be checked, and a clean result is not confirmation that your voice has not been misused elsewhere. This is a review tool, not a monitoring service.
Our outreach mission — bringing platforms on board
We believe every AI voice platform has a responsibility to verify that voices it synthesizes were enrolled with genuine consent — and to honor a rights holder's authorization decision before generating audio.
VoiceSeal is actively reaching out to major AI voice platforms to adopt the pre-synthesis check. We are offering to become the neutral, independent authorization and provenance layer that protects both creators and platforms.
When a TTS platform integrates the VoiceSeal pre-synthesis check API, they demonstrate a genuine commitment to creator consent. Their users can synthesize responsibly. Creators can license their voices with confidence. And platforms align with existing biometric and transparency laws — and proposed U.S. digital-replica legislation — as the standard for consent, documentation and provenance rises.
Our outreach posture to platforms is straightforward: integrating VoiceSeal is not an admission that cloning has occurred on your platform — it is a public signal that you take creator rights seriously. The platforms that engage early will be recognized as leaders. Those that don't will face increasing scrutiny from regulators, unions, and the voice artist community.
What integration looks like
A platform that integrates VoiceSeal can:
- Run a pre-synthesis check before generating audio from any enrolled voice
- Decline unauthorized synthesis requests based on the rights holder's decision
- Honor licensing terms and issue license tokens that authorize legitimate use
- Mark generated audio with a trusted C2PA content credential for downstream provenance
- Keep audit logs of authorization decisions
Integration moves protection to the point of synthesis. When a platform adopts the pre-synthesis check, a voice can be authorized or declined before audio is generated — rather than a rights holder discovering an unauthorized clone only after it has been published. That shift, from after-the-fact to before-the-fact, is the whole point of the outreach.
Current outreach status
We are in active or planned outreach with the following platforms and organizations. This list will be updated as relationships develop:
- LOVO AI (Genny) — API access requested. Pending response.
- Suno.ai — AI music generation platform increasingly used with voice-like audio. Outreach initiated. Suno's model can produce audio that closely resembles real voices in musical contexts — an emerging vector for voice identity misuse that we are actively working to cover.
- Coqui TTS — Open-source project. We are working directly with community maintainers on a voluntary verification model.
If you represent a platform, union, or organization and want to discuss integration or partnership, use the form below. We review all submissions personally.
What's coming next
- More platform integrations — expanding the set of AI voice platforms that adopt the pre-synthesis check, so authorization is honored at the point of synthesis on more services.
- Rightsholder verification out of Beta — maturing the human-reviewed verification workflow and the evidence records it produces.
- Private voice authorization — working with platforms to build consent-based mechanisms so a registered owner's decision governs private (unlisted) voice use, not just public library entries.
- Platform transparency score — a public rating for each platform based on their cooperation with voice-identity authorization. Updated periodically.
- EU AI Act Article 50 support — machine-readable synthetic audio marking required from August 2, 2026. VoiceSeal supports this through in-band C2PA content credentials, the standardized provenance mechanism (see our C2PA transparency page for verifiable reference files and current status).
Platforms that integrate VoiceSeal's pre-synthesis check API can be listed publicly as implementation partners. We believe transparency about who is doing the right thing is itself a form of accountability.