wachidamirul/vocalize-ai — reverse-engineered prompt

Build me a clean first version of this app called Vocalize AI. The README is basically just the default run instructions, so use the existing project name and make a sensible AI web app around helping people turn text into something they can hear and understand.

I want a simple page where someone can type or paste text, ask the AI to improve it, simplify it, rewrite it in a different tone, or prepare it to be read out loud. Show the result clearly, with loading and error states, and make it feel polished on desktop and mobile. If the browser supports it, add a button to read the result aloud.

Use the Gemini API key from the local env file like the project expects. Keep the code organized in the existing React and TypeScript structure, and make sure it runs with npm install and npm run dev. Look up current docs online if you need to.