The healing flat-illustration treatment prompt is pre-filled — low-saturation palette, clean geometric blocks, subtle film grain, 9:16 portrait output ready for your lock screen.
Below are 18 calibrated scene prompts: seaside, mountain, lake, cherry blossom, snowy night, starry sky… pick one and paste.
Free credits for new users — no credit card required
Healing flat landscape illustration is the look that has been quietly taking over Pinterest, Behance, and Xiaohongshu for the last two years: mountains, seas, street corners, and starry skies abstracted into geometric blocks, painted in a low-saturation cream-and-macaron palette, dusted with a thin layer of film grain. Nothing is happening — and that's the point. The mood pulls you in. This page packs that look into a calibrated treatment prompt and ships 18 ready-to-use scene templates, so you don't have to write the prompt from scratch or burn credits on trial-and-error.
Cream white, macaron pink, foggy blue, warm orange — every scene runs at deliberately low saturation. Easy on the eyes, friendly to lock-screen text overlays, never fights the time widget.
Mountains become triangles, seas become bands, clouds become ovals. Every element abstracts into solid color blocks and clean lines — the visual shorthand that lets modern flat illustration read in a single glance.
The treatment prompt locks in "vertical phone wallpaper, 9:16 ratio, fits lock screen and home screen." Output drops straight into Photos, set as wallpaper, done — no reframing.
A thin layer of film grain over clean geometric color blocks, plus generous negative space — the result reads like a 1970s American magazine spread. Quiet, warm, breathing.
The 18 prompts group into six categories. Don't know what you want? Pick a mood first, then choose a specific scene from the cards below.
Seaside, mountains, lakes, forests, grasslands — the kind of quiet that drags you out of the city. Made for the bad-day wallpaper that asks you to take a breath.
Spring blossoms, summer windows, autumn forests, winter snow — switch your wallpaper season by season. Cherry blossoms at the equinox, snowy village on New Year's Eve. The ritual writes itself.
Cafes, bakeries, windows, street corners — slow-city subjects. Warm palette, intimate framing — perfect for Xiaohongshu / Instagram covers and slide deck backgrounds.
Cloud trains, starry tents, moonlit forests — light, imaginative subjects. Doubles as kids' book illustration source material or bedtime-story slideshow art.
Tea fields, rice paddies, Jiangnan-style stone bridges — Eastern negative-space aesthetics. More restrained palette, flatter composition, a calmer cousin of the Western healing look.
Rainy bus stops, riverside towns, dusk streets — urban but quiet. The kind of image that makes a commuter exhale on the way home.
Each card is a fully-tuned prompt. Open the generator above, paste the matching scene description (the 9:16, low-saturation, film-grain parameters are already included), hit generate. GPT Image 2 reads Chinese and English natively — write the scene in whichever you prefer.
9:16 portrait output, generous negative space, no embedded text or watermarks — drop it straight into multiple use cases without redesigning anything.
Output starts at 1080×1920 and up. Save to camera roll, set as wallpaper, done. The negative space and low-saturation palette are designed to sit *under* a clock and notifications without competing with them.
9:16 lines up with Pinterest pins and Xiaohongshu cover cards perfectly. Type your title directly into the negative space — no cropping, no recoloring, just a one-line title before publishing.
Use as a vertical accent in a 16:9 deck, or fill an entire slide on a mobile-first deck. Body text reads cleanly on the low-saturation color blocks, and the healing mood elevates a generic corporate template instantly.
Use as a blog or newsletter cover, vlog intro card, or TikTok transition still. Color blocks are large and negative space is plentiful — overlaid logos, titles, and copy never feel cramped.
18 scene templates, the 9:16 portrait default, batch generation, and transparent credit cost — all wired into a single shortest path. Land on the page, run your first image.
The same OpenAI model that powers the rest of the site. Stable on low-saturation palettes, clean on geometric block edges, and respectful of the 9:16 ratio — which is exactly what this aesthetic relies on.
The treatment prompt is loaded into the editor on arrival. Want a different scene? Paste any of the 18 scene descriptions over it. Want warmer / cooler / more minimal? Edit the wording — the form behaves like the regular generator.
Generate 1, 2, 4, or 8 images at once. The same prompt produces compositional variants — pick the strongest one for your wallpaper, save the rest for later.
Exact credit cost shows above the generate button before you click. New users get free credits — enough to run all 18 scenes once before deciding whether to upgrade.
Output is a high-resolution PNG, ready to long-press or drag to disk. Free runs include a light watermark; paid plans output completely clean — fine for Xiaohongshu, newsletters, or client deliverables.
Common questions about the prompt templates and GPT Image 2 generation. More questions? Email support@imagesv2.ai.
Either works. GPT Image 2 understands both languages with comparable fidelity, especially for style keywords like "flat illustration," "low saturation," and "film grain." The 18 scene prompts on this page are written in clean Chinese for the source brief, but every card has an English description above — write your scene in whichever language you prefer.
9:16 is the default because the primary use case for healing landscape illustrations is phone wallpapers and Pinterest / Xiaohongshu vertical pins. For landscape, edit the prompt: replace "vertical phone wallpaper, 9:16 ratio" with "horizontal desktop wallpaper, 16:9 ratio" and the model will adjust — landscape works well for slide-deck backgrounds, desktop wallpapers, and video covers.
No. The prompt asks for "subtle film grain," so the model adds a thin noise layer that reads as texture, not noise. Want it cleaner? Remove "subtle film grain" from the prompt. Want it more retro? Replace it with "visible film grain and soft fading".
Add a sentence in front of the scene description like "a girl in a white dress standing on the grass with her back to the viewer." GPT Image 2 will fold the figure in as a flat silhouette without breaking the style. Keep the description as a back view or silhouette — recognizable faces tend to clash with the abstract aesthetic.
AI generation is stochastic — the same prompt won't produce identical images twice, but the style (low saturation, geometric blocks, grain) stays consistent. If the first take isn't right, regenerate; for tighter control, add specifics to the prompt ("sky occupies the top 60%," "cabin in the bottom-right").
Yes — switch the generator to image-to-image mode, upload your photo, and use a prompt like "redraw the attached image as a healing flat-style landscape illustration: low-saturation palette, geometric blocks, subtle grain." GPT Image 2 will keep the rough composition and re-render it as flat illustration.
Free credits for new users. The treatment prompt is pre-filled, output is 9:16 portrait — generate, save to your camera roll, set as wallpaper.








Quality Comparison (gpt-image-2)
| Quality | Speed | Image Detail | Credits/Image | Best For |
|---|---|---|---|---|
| Low | Fastest (3-8s) | Good composition, less detail | 10 | Quick iterations, bulk generation, social media |
| Medium | Moderate (10-20s) | Rich details, good textures | 40 | Marketing images, presentations |
| High | Slower (20-40s) | Highest fidelity, finest details | 110 | Print, posters, premium assets |
| Auto | Model decides | Auto-selected by model | 40 | When unsure |
Model Comparison
| Model | Highlights | Low/Image | High/Image |
|---|---|---|---|
| gpt-image-2 | Latest, best results | 10 | 110 |
Cost-Saving Tips
Enter a text prompt describing the image you want. Adjust parameters like size (Square, Landscape, Portrait), quality (Low/Medium/High), output format (PNG/JPEG/WebP), and background (Opaque or Transparent). Click "Generate" and GPT Image 2 will create a brand-new image from your description.
Tips for Better GPT Image 2 Results
Upload a source image, write a prompt describing the changes you want, and GPT Image 2 will modify the image accordingly. Without a mask, GPT Image 2 decides which areas to change. With a mask, you can precisely control which regions are modified.
You can use Edit mode for GPT Image 2 image-to-image generation without a mask. Simply upload a reference image and describe the transformation you want in the prompt — for example, "Convert this photo to a watercolor painting" or "Reimagine this scene in a cyberpunk style". GPT Image 2 will use your image as a reference to generate a new version.
Example GPT Image 2 Prompts for Image-to-Image