QR passport
Create a QR tag with contacts to print and attach to a collar.
0 active QR tags
Qaz.pet
Pet tools
Practical services for a QR passport, initial vet guidance, breed detection, and photo search.
43
photos in search index
0
active QR tags
2
active reports
Vet guidance
Describe symptoms and get an urgency guide. It does not replace an in-person vet visit.
4 messages in the vet module
Breed detection
Upload a photo to get likely breed suggestions for a pet profile.
38 open cards can be enriched with breed data
Photo search
Compare a lost or found pet photo with cards on the platform.
43 media files are available for comparison
What is already live
This page shows connected modules and the current data volume used by product workflows.
QR tags
Creation, public landing page, PNG code, and scan logging run through the Qaz.pet API.
Photo search
Photos are compared with animal and lost/found indexes through the shared embedding pipeline.
AI guidance
Vet chat and breed detection are available as first-pass helpers before an application or moderation step.
Vision layer
A photo becomes a verifiable signal
The flow is an operational pipeline: the user uploads a photo, the system shows similar cards, and the final decision stays with a human reviewer.
What we take from open-source practice
GitHub and Hugging Face research points to the most useful lightweight additions: visual search, breed hints, and a mobile operational layer for shelters.
CLIP / image similarity
A practical fit for image embeddings, similar-photo search, and human review of candidates.
Breed helper
Breed classification models are useful as profile hints, but should not replace moderator review.
Shelter operations
Open-source shelter projects converge on cards, records, mobile access, and post-adoption follow-up – this strengthens our roadmap.
Qaz.pet lab
Next level – re-identification
Instead of only guessing breed, we can compare individual animals by visual similarity: face, color, markings, pose and image quality. This strengthens lost/found, but requires a local benchmark on Qaz.pet photos first.