🐶 Paw Detective

A simple multi-agent LangChain application that identifies dog breeds from images and retrieves detailed information about the detected breed.

How It Works

1. Image Inference Agent

   • Uses a CNN model to predict the dog breed from an uploaded image.

2. Information Retrieval Agent

   • Based on the predicted breed, retrieves relevant breed-specific information from curated sources.

3. LLM Output

   • Combines the retrieved content to generate a natural language description of the breed.

Example Flow

1. Upload a dog photo.
2. Model predicts: Golden Retriever
3. System responds:

   "This is likely a Golden Retriever — a friendly, intelligent breed known for its loyalty and ease of training."

🔧 Tech Stack

• CNN (TensorFlow) MobileNetV2 architecture for breed classification
• LangChain for multi-agent orchestration
• LLM (llama) for final response generation

Features

• Lightweight and easy to extend
• RAG-style architecture using multi-agent design
• Plug-and-play image classification + retrieval

---

Feel free to clone, modify, or integrate into your own AI assistant projects!