Full prompt breakdown: turning drink photos into isometric illustrations with GPT image 2.0
op7418 · x · 2026-11-03
op7418 publishes the full prompt behind his "drink photo → isometric miniature building" GPT image 2.0 workflow:
- Photo fidelity is top priority: use the uploaded photo directly as the image-to-image reference (no edge extraction); preserve cup proportions, lid/straw structure, brand logo and colors, drink layers/toppings, collab mascots.
- Viewpoint spec: elevated three-quarter isometric view, 45° front-side, 30° down, showing roof, front, side and base depth.
- Optional fields: brand, drink name, collab info, bean origin — auto-recognized from the photo if left blank, never fabricated.
- Output: stitch original photo and illustration into one HD comparison image with code.
Related event: GPT image 2.0 turns drink photos into isometric store illustrations(2 posts)→
More from Multimodal
- Hermes Agent + MeshyAI turns conversational prompts into 3D game characters — egeberkina · 2026-11-03
- Remotion launches new AI editing experience for faster agent-made video edits — aziz4ai · 2026-11-03
- Open-source open-genjutsu swaps video performers with Seedance or Wan, undercutting $150/mo apps — matchaman11 · 2026-11-03
- MOSS image renders impress Reddit users with realism — PhoenixtheII · 2026-11-03
- Hala Launches AI Voice Studio for Arabic, Featuring ElevenLabs V4 — aziz4ai · 2026-11-03
- Midjourney SREF share: bold red, grey figure, fine black city lines — tisch_eins · 2026-11-03