Build an AI Image Generator with Claude Code

Build an AI Image Generator web app. People search for "ai image generator" about 3,000,000 times a month, plus related searches like "ai art generator", "text to image ai", "ai picture creator". The goal is a tool that ranks for those searches and makes money.

What it should do:
Build an AI image generator web app. Create a homepage with a large text input box labeled 'Describe the image you want to create' and a bright 'Generate Image' button. When users click generate, show a loading animation with progress bar. Display the generated image in a large preview area with download and regenerate options. Add style presets like 'Realistic', 'Artistic', 'Cartoon', 'Photography'. Include a gallery of recent generations. Use a modern design with gradients and glass morphism effects. Focus on the text-to-image generation flow first - you can integrate actual AI models later.

How to build it:
1. Read this folder first. If it is empty, set up Next.js (App Router, TypeScript, Tailwind). Write a short PLAN.md with the pages, components, any library or API you need, and what you are leaving for later. Then build.
2. Put the tool itself on the homepage so a visitor can use it within seconds, with no sign-up. Do the processing in the browser where you can, so it is fast and cheap to run. Look at how Midjourney, DALL-E 2, Stable Diffusion do it and make ours simpler and faster.
3. Make it rank. Use "ai image generator" in the title tag, H1 and meta description. Under the tool, add a short how-to and an FAQ with FAQPage schema, and add a sitemap and robots.txt. Scaffold these programmatic pages from one template and a data file: /ai-image-generator-for-[industry], /generate-[style]-images-with-ai, /ai-[image-type]-creator.
4. Monetisation: Freemium: 10 free images/day, $15/mo unlimited + high resolution. Add a pricing page and the free usage limit, and leave Stripe checkout as a clearly marked stub with the environment variables it needs.
5. Run it locally and test the main flow end to end, including edge cases like empty input and very large files. Fix what breaks, then finish with a summary of what works, what is stubbed and how to deploy it to Vercel.

Paste it into Claude Code in an empty folder. It plans first, so read PLAN.md and steer before it starts building.