Search papers, labs, and topics across Lattice.
VetClaw is an innovative edge-cloud multimodal system designed for early veterinary disease screening, integrating a camera module for image capture and a server-hosted vision-language model for zero-shot disease classification. By decoupling agent interaction from workflow orchestration, VetClaw enhances the diagnostic process through features like input validation, image transmission, and safety checks, ultimately enabling a more dynamic and responsive screening system. The findings reveal that while image-only predictions are limited, incorporating symptom descriptions significantly boosts classification accuracy, demonstrating the system's potential to transform veterinary diagnostics.
Symptom-guided multimodal inputs can dramatically enhance zero-shot disease classification in veterinary settings, outperforming traditional image-only approaches.
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.
