Welcome to VETeran!
Stay updated with the latest information about our project, our team and our progress.
We are senior Computer Science students at Bilkent University, Türkiye, developing our senior project VETeran: clinical-grade veterinary decision support, with photo and symptom triage grounded in trusted veterinary literature and every verdict traceable to its source.
Learn more ↓About Our Solution
The problem
Costly or uncertain visits
A worrying symptom usually means an expensive clinic trip or a confusing search, with no way to tell which one the moment actually calls for.
Contradictory advice
Forums and generic chatbots give confident answers with no way for an owner to check where they came from.
Unverified professionals
Anyone can claim to be a vet online, and owners have no reliable way to tell who actually is one.
How VETeran works
Two retrieval pipelines feed a single safety gate: cited rules set the minimum urgency, and retrieved evidence explains it.
Strict rules, searchable evidence
Species-aware symptom answers go through a strict rule lane that catches red flags and a search lane that pulls and reranks matching literature passages. Every verdict links the exact passages behind it, and no red flag is ever downgraded.
Photos checked against real cases
A vision model embeds the owner's photo and finds the closest cases in trusted veterinary literature. Low-similarity matches are flagged, never guessed, and a photo can only raise concern, never lower an urgency level.
Roadmap
Knowledge base
Source and license veterinary literature; passage-level index with vector search.
Hybrid clinical triage
Rule lane, search lane and safety gate, with citations in every verdict.
Visual symptom triage
Image embeddings, similar-case search, confidence thresholds and fallback.
Clinical validation
Veterinarian review of verdicts and a pilot with partner clinics.