Better pet health, built with vets

Better health for the patients who can't ask for it.

Animals can't tell us what's wrong — and off-the-shelf AI gets it wrong about half the time. Ruvv Labs is building the clinical data that fixes that, with the people who already know: credentialed vets. And we pay for their expertise.

Member of
NVIDIA Inception Program member badge

Ruvv Labs has been admitted to the NVIDIA Inception Program for startups.

NVIDIA and NVIDIA Inception are trademarks and/or registered trademarks of NVIDIA Corporation in the U.S. and other countries.

The problem

Pet owners increasingly lean on AI for their animals' health. But most of it is trained on images scraped from the web or labeled by non-experts — with no clinical ground truth. That is how you get tools that are confidently wrong.

~52%
In our own benchmark of ~10,000 vet-labeled clinical images, off-the-shelf AI disagreed with the veterinarians on roughly half of its reads.

Better tools for pets need better data — and better data needs real vets. That is what we are building — with the profession, not around it.

What we build

A large, expert-labeled library of real clinical scans — captured and labeled by credentialed veterinarians on a controlled medical vocabulary, not free text.

  • 10 scan types — dental, ears, eyes, skin & coat, body condition, orthopedic gait, and more.
  • 600+ clinical features drawn from recognized veterinary standards — AVDC dental charting, WSAVA body condition, CAFSS dermatology, LOAD, and AAHA guidelines.
  • Every scan carries both a veterinarian's read and an AI read, reconciled against each other.
How we keep it honest

Every submission runs an AI-plus-human verification pipeline: quality scoring, species and body-part matching, biometric duplicate and fraud detection, gold-standard honeypots, and continuous random audits.

Around 97% of automated decisions agree with expert human review — and nothing is silently trusted.

The dataset today
627
credentialed vets accepted across 35 countries — and growing
22K+
vet-labeled clinical photos & videos, and growing
226
breeds represented
10
clinical scan types
600+
features on a controlled medical vocabulary
For veterinarians

We are early, and we are building this with the profession. If you are a qualified vet, the images you already take during an exam are worth something — contribute your clinical judgment, and be paid for it. Remote, flexible, paid in your local currency via trusted international payment providers.

Request access See how contributing works →
Built responsibly

Contributors submit only what they choose to — with owner permission and no identifying client information. Every submission is reviewed individually, and the data is handled with clinical-grade care.

Contact

For partnerships, research, or general enquiries.

[email protected]