GPT image 2.0 demo: one prompt turns coffee photos into isometric shop illustrations
op7418 · x · 2026-11-03
op7418 shares a GPT image 2.0 prompt that turns a real coffee/tea photo into a chunky-outline isometric miniature building illustration. In his example, a Grid Coffee Americano became a cup-shaped storefront with Yunnan Qiupo Estate coffee details and tiny staff characters.
Key technique: use the original photo as the visual reference for image-to-image generation, preserving the cup silhouette, lid, logo and brand colors while redesigning the building functions, then stitch original and generated images side by side with code.
Related event: GPT image 2.0 turns drink photos into isometric store illustrations(2 posts)→
More from Multimodal
- Synthesia launches ChatGPT plugin that turns docs and decks into branded videos — alexvoica · 2026-11-03
- HunyuanVideo 1080p added 50 minutes per clip on a 4090 — sysmem fallback strikes again — Suspicious_Pizza9529 · 2026-11-03
- RL for Diffusion Models: A Deep Dive on Why Rewards Beat DPO Pairs — kastnerkyle · 2026-11-03
- A 17-second AI car ad built story-first: script, then storyboard, then generation — EdenMakers-ai · 2026-11-03
- CapCut Multi-Track Workflow for AI Fashion Try-On Videos: Sync Outfit Changes to the Washing Drum — Aiden_Tech_Ai · 2026-11-03
- ComfyUI Workflow Generates 1080p Video with MiniMax H3 on Only 6GB VRAM — cgpixel23 · 2026-11-03