VetClaw: An Edge-Cloud Multimodal Agentic System for Veterinary Disease Screening
2026-07-28 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionMachine Learning
AI summaryⓘ
The authors created VetClaw, a system that helps spot diseases in animals early by using a camera and smart software. It sends pictures and symptom info from the animal to a cloud model that guesses possible diseases without prior training on them. The system separates tasks between devices at the edge and the cloud, allowing it to manage workflows, check safety, and handle errors better than simple image classification. They found that adding symptom details improves disease detection compared to just using images. Overall, the authors designed VetClaw to be a smarter and safer tool that can manage complex processes and ask for help when unsure.
edge computingcloud computingmultimodal learningvision-language modelzero-shot classificationworkflow orchestrationveterinary disease screeningdiagnostic supportimage classificationmodel safety
Authors
Syed Mhamudul Hasan, Anas AlSobeh, Hussein Zangoti, Abdur R. Shahid
Abstract
We present VetClaw, an edge-cloud multimodal agentic system for early veterinary disease screening. VetClaw uses a camera module as an edge sensing device and sends captured images, together with optional symptom descriptions, to a server-hosted vision-language model for zero-shot disease classification. The system separates agent interaction from workflow orchestration: OpenClaw provides scheduling, tool access, user interaction, and notification services on the edge device, while LangGraph manages the stateful screening workflow, including input validation, image transmission, model invocation, safety checks, conditional routing, failure handling, and structured logging. This design moves beyond static image classification by enabling the system to collect visual evidence, invoke external models, apply deterministic safety rules, and generate diagnostic-support alerts. Results show that image-only VLM prediction remains limited, whereas symptom-guided and multimodal inputs improve zero-shot classification performance. Thus, VetClaw transforms a static prediction model into a coordinated, safety-aware system that can use tools, manage workflows, handle failures, and escalate uncertain cases.