How it works
How presence becomes proof
One proof: attest at capture, carry with the content, verify anywhere.
Attest
A companion sensor measures the body producing the speech: radar for contactless devices, bone conduction for worn ones. That measurement is cryptographically bound to the audio and signed on the device at the moment of capture. New sensor types can join without changing the proof.
Carry
The proof is bound to what was said, not to one audio file. It follows the words into a transcript, an email or a quote, wherever the tools in between preserve it. We sign into whatever provenance standard the ecosystem converges on. Today that is C2PA. We are the missing first link: provenance should start at the larynx, not the file.
Verify
Anyone downstream can check the proof without asking us: a browser, a bank, a meeting app. The same check can run on a device before a command executes.
OUR APPROACH
The proof is made where the voice is made
At creation
Not after the fact — no arms race to lose.
On the device
Not in the cloud. Only the proof travels.
On the content
Not the account. A verified account can still post a fake.
Blogs
Two levels of proof
Verified Human
A live person is speaking at the microphone.
Trusted Human
That person is the enrolled user, on a registered device. A pseudonym works.Neither needs a camera or an iris scan.
Blogs
Insights
One argument, unpacked one post at a time — why ground truth is collapsing, why detection can't keep up, and what proof of human presence means for every voice, call, and command.
04 — HARDWARE
How you get it
TODAY
Standalone USB-C sensor
Earbud
NEXT
Same sensor, built in
