Authenticity on every claim photo

We attempt to verify any photo. We gives a definitive answer for photos captured via our SDK in vendor apps. Newer devices with hardware attestation, such as the Google Pixel 10+, Apple iPhone 18 and Canon EOS R1, are supported by default.
All other images go through our advanced predictor without our false positive guarantee

They still work, but with a confidence score. A photo that wasn't certified at capture is marked as unsigned and goes through our time and location intelligence detector before landing in your normal adjuster review, just as it would today. As more phones ship with built-in signing, that review queue shrinks.

Yes. Our patented combination of AI and perceptual hashing recovers an Autograph from screenshots and other types of image manipulation.

Yes. Every Autograph contains an encrypted capture time, signed by the phone's secure chip at the moment the shutter fires. Otherwise, we use key details within the photograph, such as shadows.

Yes. Every Autograph contains encrypted, precise location details, signed by the phone's secure chip at the moment the shutter fires. Otherwise, we use key details within the photograph, such as landmarks and weather.

Autograph's use LiDAR to generate a depth field with the photo. We provide an extra score of photographs of a screen or a printout, since it shows up as a flat surface instead of a real scene

The Autograph can be recovered from compressed or resized photographs. If a photo is edited, the verifier can show that it changed after capture and show the source photograph.

Yes. We extract photos embedded in PDFs and inspection reports and checks whether they are certified, even when the PDF carries no image metadata. It applies the same perceptual hashing used for our Autographs to match each embedded image back to its original certified capture. This means a report with thousands of images can be reviewed as a batch, rather than photo by photo.

Not yet. Today we can Autograph photos and documents. Video is on our roadmap.

An Autograph is an encrypted digital certificate the phone's secure chip attaches to a photo the moment it's taken. It records the image, time and location, and any change afterwards breaks it. Anyone with the photo can check it in about a second, even in airplane mode, since it is saved to an open source bucket for public, auditable verification.

The two approaches complement each other. We certify the photos taken through the app, and AI-detection tools remain useful for the rest.

An Autograph means the photo or document is authentic. We don't give a percentage score, because verification is a cryptographic check that a photo either passes or fails. There is no threshold or confidence level: we are sure the photo came from a genuine device.

Soon. Whether a certified photo is admissible depends on the jurisdiction and the specific case, and our counsel is reviewing the question now. The signature is a real X.509 digital certificate created in the phone's secure hardware at capture, which is the kind of record US rules such as FRE 902(13) allow to be authenticated by certification rather than live testimony.

No. Certificates are created with standard public-key cryptography. There's no per-photo blockchain cost nor wait.

No. Photos and location go directly from the claimant's phone to the carrier, encrypted end to end. We only receives a cryptographic fingerprint of the image, which can't be turned back into the photo.

No. Verification itself uses no generative AI. It's a cryptographic check. Optional add-ons such as damage estimates use offline ML/AI models, and never use your data for training.

Only the claimant's device. Each key is generated inside the device's secure chip and can never leave it.

Not yet. Revocation is planned, with revoked keys published to an open registry.

Ed25519 and ECDSA-P256 signatures, SHA-256 hashing, and Meta's PDQ perceptual hashing.

We connect to your claims system in several ways, and you can use one or combine them.

• SDK in a mobile app: added to the app your claimants already use, so the phone's secure element signs each photo at capture.
• REST API at intake: your claims system sends each photo to our verification API when it arrives. Valid photos proceed automatically, and missing or invalid signatures are flagged with an evidence bundle.
• Browser extension for adjusters: lets adjusters check photos in their browser.
• Prebuilt P&C connectors: connects to claims platforms without custom integration.
• Platform partners: available through partners.

Yes. The photo is certified on the phone the instant it's taken, and checking it at intake takes about a second.

Our focus is building a more fraud-resistant claims pipeline for insurance.
In the future, we believe Autograph's can be valuable in other regulated industries!

Our paid design partnership is built around integrating our proof of concept, to ensure it hits your agreed-upon success criteria.

Not today. Damage estimation is on our roadmap. What we ship now is verifying that a photo is authentic: signed at capture by the phone's secure hardware, then checked at intake. The approach we're building runs local AI models on photos after they have been signed, and can use a claim's full album plus the VIN to make estimates more precise. The auto-claim demo in our deck that estimates damage and checks a repair bill is illustrative, not a shipping product.