Fake Gemini chats, drawn pixel by pixel
You’re writing a post about prompt engineering and you need a screenshot of a Gemini conversation. Opening the real app, logging in, staging the exact prompt and reply, cropping out the sidebar clutter, that’s ten minutes gone. Or you type the turns here and download a clean PNG in about thirty seconds.
This mockup doesn’t screenshot anything. It draws the Gemini interface straight onto an HTML5 canvas: your prompt sits in the soft blue-gray bubble on the right, and Gemini’s reply flows on the left next to the four-point sparkle rendered with its blue-to-purple-to-pink gradient. Every pixel is generated in your browser.
What you can change
Add as many turns as you want. Each one is either a prompt (right bubble) or a Gemini response (left, plain text with the sparkle). Rewrite them inline, shuffle the order with the up and down buttons, delete the ones you don’t need.
- The Gemini wordmark header toggles on or off, with the same gradient the real product uses.
- Dark mode flips the whole canvas to Gemini’s near-black background with light text.
- Multi-paragraph replies work. Press Enter for line breaks, leave a blank line between paragraphs, write numbered lists, and the wrapping keeps up.
The reply text wraps exactly the way it’ll export, so the preview is honest. No surprises when you hit download.
Who reaches for this
Course creators building slides about AI assistants. They need consistent visuals, not fifteen mismatched real screenshots with someone’s account name in the corner.
Then there’s the meme crowd. A well-timed fake Gemini answer is comedy gold, and this makes one in under a minute. Product designers use it too, mocking an assistant flow into a prototype before a single line of the real thing exists.
Honestly, anyone who’s tired of blurring out their email address in a real capture.
Getting the details right
Real Gemini replies have a rhythm. They open with a short framing line, break into a numbered or bulleted list, and often close with a follow-up question like “Want me to expand on any of these?” Copy that structure and the mockup reads as genuine.
Keep prompts short. One or two sentences. Long walls of user text look staged, because real people don’t type essays into the prompt box. Match the theme to where the image is headed too: dark mode for a dark slide deck or a video, light mode for a printed handout.
Common questions
Does this use my real Gemini account?
Nope. There’s no login, no API call, nothing leaves your machine. The canvas is drawn locally from the text you type.
Can I export the dark theme?
Yep. Flip the Dark Mode toggle and the whole image switches to Gemini’s dark background, bubble, and text colors before you download.
How sharp is the PNG?
The canvas renders at 2x, so a 560-pixel-wide design exports around 1120 pixels across. Crisp on retina screens, fine for print.
Why does the sparkle look like the real one?
Because it’s the actual four-point star shape, drawn as a vector path and filled with Gemini’s blue-purple-pink gradient. It scales without blurring at any size.
Can I reorder messages after typing them?
Each turn has up and down arrows. Rearrange them however you like and the canvas redraws instantly.