Run and deploy your AI Studio app

This contains everything you need to run your app locally.

View your app in AI Studio: https://ai.studio/apps/drive/1V0rn7pwslaxnFRTYfBazPPHU4RVq-Jr8

Run Locally

Prerequisites: Node.js

1. Install dependencies: npm install

2. Add your API keys to .env (see .env.example). Keys are only used by the API server, never by the frontend.

3. Start both the API and the app in one go:

   npm run dev:all

   This runs the API server (port 3001) and Vite (port 3000) together. The app proxies /api to the server.

   Alternatively, use two terminals:

   • Terminal 1: npm run server (API on 3001)
   • Terminal 2: npm run dev (Vite on 3000)

Deploy on Vercel

On Vercel, the middleware is implemented as serverless functions, not a long-lived Express server:

• Each file under api/ becomes a separate serverless endpoint:
  • api/plot-details.ts → POST /api/plot-details
  • api/refine-plot.ts → POST /api/refine-plot
  • api/characters.ts → POST /api/characters
  • api/next-episode.ts → POST /api/next-episode
• The frontend (Vite build in dist/) is served as static files. The vercel.json rewrites all non-/api routes to index.html for client-side routing.
• Environment variables (e.g. OPENROUTER_API_KEY, AI_PROVIDER) must be set in the Vercel project: Settings → Environment Variables. They are only available in the serverless functions, so API keys never reach the browser.
• Deploy with the Vercel CLI (vercel) or by connecting your repo in the Vercel dashboard. No need to run the Express server/ on Vercel; the api/ routes replace it.