# Neeraj Aggarwal Digital Twin A production-ready personal portfolio with a server-side AI Digital Twin chat. ## Structure ```text . ├── About_Me.pdf # Source biography, loaded by /api/about ├── public/ │ ├── index.html │ ├── styles.css │ └── app.js ├── server.js ├── package.json ├── .env └── .env.example ``` ## Run locally 1. Keep `About_Me.pdf` in the project root. The browser loads the exact extracted text through `GET /api/about`. 2. Ensure `.env` contains the requested variable: ```env OPENAI_API_KEY=your_openrouter_api_key OPENAI_BASE_URL=https://openrouter.ai/api/v1 OPENAI_MODEL=openai/gpt-4o-mini PORT=3000 SITE_URL=http://localhost:3000 ``` The server also accepts the legacy `OPENAI-API_KEY` spelling currently present in this workspace, but `OPENAI_API_KEY` is recommended. 3. Install and start: ```bash npm install npm start ``` Open `http://localhost:3000`. ## Deployment notes - Keep `.env` out of version control; never expose the API key to browser JavaScript. - Set environment variables in the hosting provider's secret/configuration panel. - Set `SITE_URL` to the deployed origin so OpenRouter receives the correct HTTP referer. - The app uses Node.js 18.17+ for native `fetch`. - For a different OpenAI-compatible router, set `OPENAI_BASE_URL` to its `/v1` endpoint and update `OPENAI_MODEL`. # ai_learning