NirdeshakAI => https://nirdeshak-ai.vercel.app/
An intelligent multilingual assistant designed to help users understand and access Indian government services easily. Built with FastAPI backend and React frontend with voice capabilities.
- Multilingual support for Indian regional languages, Hindi, and English
- Voice input/output using speech recognition and text-to-speech
- Clean, responsive UI with real-time chat interface
- Markdown formatting for structured responses
- Direct links to official government websites
- Python 3.8 or higher
- Node.js 16 or higher
- npm or yarn package manager
Create a .env file in the server directory with the following variables:
AZURE_OPENAI_API_KEY=your_azure_openai_api_key
AZURE_OPENAI_ENDPOINT=your_azure_openai_endpoint
SARVAM_API_KEY=your_sarvam_api_key- Navigate to the server directory:
cd server- Create and activate a virtual environment:
python -m venv myenv
source myenv/bin/activate # On Windows: myenv\Scripts\activate- Install dependencies:
pip install -r requirements.txt- Start the FastAPI server:
uvicorn main:app --reload --port 8000The backend will be available at http://localhost:8000
- Navigate to the client directory:
cd client- Install dependencies:
npm install-
Update API configuration (if using local backend):
- Open
src/api.jsx - Change
BASE_URLto"http://localhost:8000"
- Open
-
Start the development server:
npm run devThe frontend will be available at http://localhost:5173
cd client
npm run buildcd server
uvicorn main:app --host 0.0.0.0 --port 8000POST /query- Submit user queries and get AI responsesPOST /transcribe- Convert audio to textPOST /tts- Convert text to speechGET /docs- API documentation (FastAPI auto-generated)
- Backend: FastAPI, Python, Uvicorn, Azure OpenAI, SarvamAI
- Frontend: React, Vite, TailwindCSS, Framer Motion
- Speech: Web Speech API, SarvamAI Speech Services
The application supports hot reloading in development mode. Make changes to the code and see them reflected immediately.
This project is developed for hackathon submission.