An advanced, multimodal AI assistant that acts as a medical consultant. This project integrates Computer Vision, Speech-to-Text, Large Language Models (LLM), and Text-to-Speech to create a seamless "Doctor-Patient" interaction.
The AI can "see" medical images (like X-rays or symptoms), "hear" patient queries, analyze the data using medical logic, and "speak" back with a realistic human voice.
🗣️ Voice Interaction: Talk to the AI naturally; no typing required.
👁️ Vision Capabilities: Upload medical images for analysis (skin issues, X-rays, reports).
⚡ Ultra-Fast Inference: Powered by Groq for near-instantaneous LLM responses.
🧠 Medical Persona: Custom system prompts designed to mimic a professional, empathetic doctor.
🔊 Realistic Voice: Uses ElevenLabs for high-quality, lifelike speech output.
🖥️ Interactive UI: Clean web interface built with Gradio.
Python: Core programming language.
Groq API: Runs Llama 3 (Logic) and Whisper (Transcription) with extreme speed.
ElevenLabs: Generates realistic AI speech.
Gradio: Front-end user interface.
RAG Implementation: Connecting the bot to a Vector Database (Pinecone/ChromaDB) loaded with real medical textbooks for higher accuracy.
History Memory: Allowing the bot to remember previous turns in the conversation.
Deployment: Hosting the app on AWS or HuggingFace Spaces.
This project is for educational purposes only. The AI is not a licensed medical professional and should not be used for real medical emergencies or diagnoses.
Vijay Sonavane - Ai Engineer
FFmpeg: Handles audio recording and playback processing.
SpeechRecognition & Pydub: Audio capture and manipulation.