If you create food content, you know a beautiful, practical recipe card often beats long copy—especially Xiaohongshu-style step tutorials: appetizing, clearly laid out, and human.
It used to mean shooting (lighting, styling, plating, retouching) or hiring a designer. GPT image2 makes this shockingly simple.
I used one structured prompt to generate tutorial images for two home-style dishes:
- 👉 Braised pork ribs (Hong Shao Pai Gu style)
- 👉 Stir-fried pork with peppers (La Jiao Chao Rou)
Good enough for notes or product detail pages.
1. Prompt template (copy-paste ready)
GPT image2 does not need a novel. My structure:
Create a Xiaohongshu-style viral cooking tutorial image (3:4 vertical).
Dish: 【菜系 / DISH】
Mood: cozy home cooking vibe, warm, inviting, slightly aesthetic lifestyle feel.
Composition: 4–6 step grid layout, clean spacing, easy to scan.
Visual style: soft natural lighting, slight film tone, warm color grading, realistic food photography.
Steps: 【步骤 / STEPS】
Details: include hands interacting with food, show textures (oil, steam, sauce), slight motion feeling, add small Chinese annotations like "切块", "搅拌", "下锅", "翻炒".
Avoid: overcrowded layout, too much text, unrealistic AI artifacts.
Replace 【菜系 / DISH】 and 【步骤 / STEPS】 with your real dish and steps—AI handles the rest.
2. Case 1: Braised pork ribs
I filled in:
Rätt：Braised pork ribs—tender, glossy sauce, perfect over rice
Steg：Chop ribs → blanch in cold water with ginger & cooking wine → caramelize sugar (rock sugar + little oil) → coat ribs → sauté aromatics (scallion, ginger, star anise, bay) → soy sauces & wine → cover with hot water → simmer 40 min low → reduce high heat, scallion garnish
GPT image2 delivered a 3:4 vertical with ~9-step grid, real hands (“stir”, “garnish”), thick sauce and steam, labels like “blanch”, “caramelize”, “season”—late-night comfort-food note vibes.
3. Case 2: Stir-fried pork with peppers
Input:
Rätt：Stir-fried pork with peppers—spicy, savory, homestyle
Steg：Slice pork belly → cut green chilies → mince garlic & bird’s eye chili → dry-fry pork to render fat → garlic & chili → high-heat peppers → soy, oyster, salt, MSG → toss and plate
Results: realistic fat sheen on pork, vibrant green peppers, precise Chinese tags (“render fat”, “fragrant”, “wok”), warm kitchen tone—no cheap plastic AI look, like a viral phone tutorial.
4. Value for food & lifestyle creators
| Scenario | Traditional | GPT image2 |
|---|---|---|
| Xiaohongshu / Douyin recipe cards | Shoot + retouch + layout, 2h+ | ~5 min, easy tweaks |
| Multi-step tutorials | Miss steps, slow layout | Auto grid, clear logic |
| Consistent style | Unstable light/background | Same prompt, same look |
| Swap ingredients/steps | Reshoot | Edit text, regenerate |
Home cooking, baking, meal prep, regional snacks—GPT image2 can be your 24/7 food photographer.
5. Closing: AI images are rewriting content production
Whether you run social (Xiaohongshu, Douyin, WeChat), e-commerce (meal kits, sauces, cookware), or recipe apps—batch-ready tutorial images matter.
- Social & short video
- E-commerce SKUs
- Recipe sites / apps
GPT image2 lowers the bar to “if you can type, you can ship.” Multilingual prompts work for Pinterest, Instagram Reels, too.
6. Quick advice
Do not treat it as full replacement for shooting—use it as an idea accelerator and efficiency tool:
- Test layouts with AI first
- Shoot hero content from winning styles
- Or use AI for daily updates
Try your signature dish in GPT image2 once—you may thank yourself later.
Nyckelord: GPT image2 / Xiaohongshu recipe / food tutorial / AI images
Tags: #GPTimage2 #Xiaohongshu #FoodTutorial #AIImages
Disclaimer: personal testing; results vary by prompt and dish. Follow platform rules for publishing.
