AI Post

Ok so I hate working with Figma and spending time on grunt work so I decided to try something last weekI used Puppeteer to take screenshots of all my SaaS and client projects we done at AstroMVP and then send them to both Nanao Banana 2 Setup Google API key for Nano Banana from Google AI studio and ask Claude code to setup Puppeteer or Chrome MCP to take screenshots and then send each image 1 by 1 to Nano Banana with a prompt like below:Present this screenshot inside a browser window. Follow these rules EXACTLY: 1. The browser frame is a thin light gray (#e5e5e5) top bar, about 40px tall, with 3 small colored dots (red, yellow, green) on the left side only. No URL bar, no tabs, no other elements. 2. Below the top bar, the screenshot fills the ENTIRE remaining space edge to edge. No inner padding, no inner margins. 3. The browser window must fill at least 95% of the total image area. Only a tiny sliver of white background visible around the edges (about 2-3% on each side). 4. The browser window has very subtle rounded corners (about 8px radius) and a faint drop shadow. 5. Pure white (#ffffff) background behind the browser window. 6. The screenshot content must stay sharp, readable, and UNMODIFIED. 7. CRITICAL: The browser frame must be EXACTLY the same size, shape, and proportions in every image. Do not vary it. 8. Output 3:2 landscape ratio at maximum resolution. 9. No laptop body, no desk, no 3D perspective. Flat, front-facing only. For mobile apps (phone frame): Present this mobile app screenshot inside a smartphone. Follow these rules EXACTLY: 1. Use a minimal, modern black phone frame with rounded corners and a small notch at the top. 2. The phone is perfectly centered vertically and horizontally on a pure white background. 3. The screenshot fills the entire phone screen with no gaps. 4. The phone should be large - taking up about 85-90% of the image height. 5. Pure white (#ffffff) background. 6. No shadow, no desk, no hands, no 3D angle. Flat, front-facing. 7. The screenshot content must stay sharp, readable, and UNMODIFIED. 8. Output 3:2 landscape ratio with the phone centered.And that's it.. you can sip your coffee/tea while this is getting done or doing some other work instead of manually doing it The result is 80% okay-ish with future models it'll be hard to recognize if it's manually done or using AI.Good luck! Posted Mar 28, 2026 at 2:28PM