VetAgent - Multi-Agent Veterinary Assistant

A conversational AI veterinary assistant built for the ElevenLabs Global Hackathon. VetAgent uses multiple specialized agents working together to provide comprehensive pet health guidance.

🎯 Project Overview

VetAgent is a voice-enabled, multi-agent system that helps pet owners understand symptoms, get care guidance, and identify when emergency veterinary care is needed. The system uses multiple specialized agents orchestrated directly in the backend.

🤖 Multi-Agent Architecture

1. Triage Agent - Initial symptom assessment and routing
2. Care Agent - Provides home care guidance and general advice
3. Emergency Agent - Identifies urgent situations requiring immediate veterinary attention
4. Main Assistant - Coordinates between agents and provides conversational interface

🛠️ Partner Technology Integration

• ElevenLabs - Voice-to-voice conversational interface
• Clerk - User authentication and session management
• OpenAI - Agent intelligence (GPT-4)
• Bolt - Backend API and data handling (optional)
• CodeRabbit - Code review and quality assurance

🚀 Quick Start

Prerequisites

• Node.js 18+
• ElevenLabs API key
• Clerk API keys
• OpenAI API key

Installation

# Install dependencies
npm install
cd frontend && npm install && cd ..

# Setup horse model (extracts from realistic-horse.zip or copies from source)
./scripts/setup-horse-model.sh

# Set up environment variables
# Create .env file with your API keys (see Environment Variables section)

# Start the backend (Terminal 1)
npm run backend

# Start the frontend (Terminal 2)
cd frontend && npm run dev

Environment Variables

ELEVENLABS_API_KEY=sk_c520685c58c6bf43d69b60255e44c28143dda1e467ac1734
ELEVENLABS_VOICE_ID=your_voice_id
ELEVENLABS_AGENT_ID=agent_8701kc6j0qzkeejrpzm39wz0m341
CLERK_PUBLISHABLE_KEY=your_key_here
CLERK_SECRET_KEY=your_key_here
OPENAI_API_KEY=your_openai_key
OPENAI_MODEL=gpt-4

📁 Project Structure

vet-agent/
├── backend/           # Express.js API with agent orchestration
├── frontend/          # Voice-enabled web interface
├── agents/            # Agent implementations (Python - reference)
├── scripts/          # Setup and utility scripts
└── docs/             # Documentation

🎤 Voice Interface

The system uses ElevenLabs Conversational AI for:

• Real-time voice-to-voice conversations
• Natural voice responses via embedded widget
• Multi-language support
• Emotional tone adaptation
• Integrated conversational agent widget

🐴 3D Horse Model

The interface includes an interactive 3D horse model:

• Realistic horse GLB model displayed using Three.js
• Interactive 3D viewer with rotation and view controls (Left/Right/Front/Back)
• Clickable condition markers (green/yellow/red) for medical issues
• Visual representation in the anatomy panel
• Model extracted from realistic-horse.zip or copied from source directory

Setup: Run ./scripts/setup-horse-model.sh to extract the model to frontend/assets/models/

🔐 Authentication

Clerk handles:

• User sign-up and login
• Session management
• User profile storage
• Secure API access

🔄 Multi-Agent Orchestration

Backend agent system handles:

• Triage agent routes cases based on symptom severity
• Care agent provides home care guidance
• Emergency agent handles urgent situations
• Direct OpenAI integration for agent responses

📊 Demo

[Link to demo video - to be added]

🏆 Hackathon Submission

Built for ElevenLabs Global Hackathon - December 11, 2024

Scoring Criteria Met:

• ✅ Clear multi-agent implementation
• ✅ Multiple partner technologies integrated
• ✅ Conversational agent theme
• ✅ Functional demo with voice interface

📝 License

MIT License - Built during hackathon event