Vidhi.AI

A web-based chat interface to interact with Vidhi.ai FastAPI backend and a React frontend. Basically, this is a RAG-tuned AI model that can answer all questions regarding The Indian Constitution. This project supports streaming responses from Ollama and can be hosted on a personal Windows machine, accessible publicly.

---

Features

• Stream responses from Ollama in real-time.
• Modern chat UI built with React.
• Fully CORS-enabled for frontend-backend communication.
• Can be deployed on a public IP without ngrok.
• Simple setup for Windows with FastAPI and Python.

---

Project Structure

.
├── ollama_api.py          # FastAPI backend serving Ollama API & React frontend
├── frontend/              # React frontend
│   ├── public/
│   ├── src/
│   └── build/            # Production build after `npm run build`
├── requirements.txt      # Python dependencies
└── README.md

---

Prerequisites

• Windows machine with Ollama installed and running locally.
• Python 3.10+
• Node.js & npm (for React frontend)
• Internet access if hosting publicly.

---

Setup

1. Backend Setup

1. Create a Python virtual environment:

python -m venv venv
.\venv\Scripts\activate

2. Install dependencies:

pip install fastapi uvicorn requests aiofiles

3. Make sure Ollama is running locally (default port 11434).

---

2. React Frontend Setup

1. Navigate to the frontend folder:

cd frontend

2. Install dependencies:

npm install

3. Build the frontend for production:

npm run build

This generates the frontend/build/ folder served by FastAPI.

---

3. Serve React via FastAPI

Modify your FastAPI app to serve the frontend:

from fastapi.staticfiles import StaticFiles

app.mount("/", StaticFiles(directory="frontend/build", html=True), name="frontend")

---

4. Run FastAPI Backend

Run FastAPI on port 8888:

uvicorn ollama_api:app --host 0.0.0.0 --port 8888 --reload

---

5. Firewall & Router Configuration

1. Open Windows Firewall:

   • Allow inbound TCP traffic on port 8888.

2. Router Port Forwarding:

   • Forward external port 8888 to your PC's local IP (e.g., 192.168.1.5).

3. Access your frontend:

http://<PUBLIC_IP>:8888/

---

6. Update React API URLs

In frontend/src/api/ollama.js:

const API_BASE = "http://<PUBLIC_IP>:8888"; // replace with your public IP

• Rebuild React:

npm run build

• Now the frontend communicates with the backend over your public IP.

---

Usage

1. Visit your public URL:

http://<PUBLIC_IP>:8888/

2. Type a message and send it.
3. Responses are streamed in real-time from the Ollama model (llama2-uncensored by default).

---

Security Considerations

• Anyone with access to the public IP can send requests to your backend.
• Recommended: Add API key authentication for /generate_stream.
• Use HTTPS (via Nginx/Caddy) if exposing to the internet.
• Keep Windows firewall and router passwords secure.

---

Dependencies

Python:

• fastapi
• uvicorn
• requests
• aiofiles

React:

• react
• react-dom
• react-scripts

---

Optional Enhancements

• Add multiple Ollama models in the frontend.
• Implement streaming updates in React (character-by-character).
• Add user authentication and API keys.
• Use Docker for easier deployment.
• Enable HTTPS via reverse proxy (recommended for public servers).

---

License

Open-source project for personal or educational use.