ElaMath is an intelligent, voice-enabled multimodal math assistant designed to help students solve and understand math problems. Simply speak your math question and optionally upload an image—such as a diagram, handwritten notes, or textbook screenshot—and ElaMath will provide clear, step-by-step explanations both in text and spoken audio
- Multimodal LLM:
meta-llama/llama-4-scout-17b-16e-instructvia Groq API - Speech Recognition: Whisper (OpenAI)
- Text-to-Speech: gTTS & ElevenLabs
- ⚡ Groq API – High-speed inference for LLaMA models
- 🧠 Meta LLaMA-4 Vision Model – Multimodal math question answering
- 🎙️ Whisper – Converts voice to text
- 🔊 gTTS & ElevenLabs – Generate natural audio output
- 🧪 Gradio – Fast web interface for interaction
- 🎙️ Voice Input: Speak your math question naturally
- 🖼️ Image Analysis: Upload diagrams, equations, or handwritten problems
- 🧠 Multimodal Reasoning: Combines voice + image to give intelligent, relevant answers
- 💬 Text + Voice Output: See and hear the explanation clearly
- 🌐 Web Interface: Lightweight, user-friendly frontend powered by Gradio
ElaMath/ ├── gradio_app.py # Main Gradio interface ├── brain_of_the_elaMath.py # Handles multimodal LLM image + text analysis ├── voice_of_the_user.py # Speech-to-text logic using Whisper ├── voice_of_the_math_instructor.py # Text-to-speech logic using ElevenLabs/gTTS ├── .env # API keys stored securely here
git clone https://github.com/iamafridi/elaMath.git cd elaMath
Create a .env file and add the following:
GROQ_API_KEY=your_groq_api_key ELEVENLABS_API_KEY=your_elevenlabs_api_key
- Run the Application python gradio_app.py
Upload an image of a geometry problem and ask:
🗣️ “What’s the area of this triangle?”
📢 ElaMath will analyze the image and your voice, then respond with a step-by-step explanation in both text and audio.
MIT License
Afridi Akbar Ifty
GitHub: @iamafridi
Portfolio: iamafrididev.netlify.app
LinkedIn: linkedin.com/in/iamafridi