Raimi's AI Art Lab

Visitors build a prompt step by step with the museum character Raimi, choosing what they want to draw, and an AI draws it. It ran on the museum floor from June to August 2026, and the app's own backend recorded 14,328 generated images: 3,049 from 17 June, 9,865 in July and 1,414 in August. One image is one successful generation; blocked or failed requests are not counted.

The app replaced an off-the-shelf GPT service that would stop responding mid-use and leave the station unusable until staff stepped in. It has not frozen once since the switch. The 10-step (+5 advanced) choice-based prompt builder shows young visitors how a prompt is put together while they play. Every image is auto-watermarked with the museum logo, and visitors take their pictures home right away by QR code. Because it runs as a web app, the kiosk could be locked to a single full-screen browser.