TikTok / Reels vertical cover with bold hook text

TikTok 封面 / Instagram Reels cover / 抖音视频封面

Sample output · gpt-image-2 · low quality (free tier)

The prompt

A vertical 9:16 video cover image. Background: a candid photograph of [SCENE], slightly
desaturated for text readability, with a dark gradient overlay from top (40% opacity)
fading to clear in the lower third. Centered upper half: the hook text "[HOOK_TEXT]"
in 130pt extra-bold sans-serif (similar to Druk or BebasNeue Heavy), white with a
thin black outline (3pt) for legibility on any background. Each line of text fits the
9:16 width with no wrapping. Below the headline, smaller secondary text "[SUBTEXT]"
at 50pt, same font, in [ACCENT_COLOR]. Bottom right: a small circular play icon ▶
hinting at video. Aesthetic: TikTok-native, high contrast, scroll-stopping.

Run it now (free)

1 free generation, no signup. Sign in for 2 · ⭐ Star for 3.

Why this prompt works on GPT Image 2

This prompt is engineered for creators on Xiaohongshu (RedNote), Instagram, TikTok, Pinterest. The phrasing prioritizes survive the 0.3-second scroll-stop test and earn a save / repin — every word in the prompt is doing one of three jobs: setting subject, controlling lighting / camera, or constraining style. What makes it land specifically on gpt-image-2: the model parses noun-clause phrasing (e.g. "centered on pure white seamless background") more reliably than vague style adjectives ("clean look"). The version above leans heavily on noun clauses — this is intentional. The Chinese "小红书 (Xiaohongshu)" cover convention is dramatically different from Western Instagram aesthetics — explicit aspect-ratio and font choice is critical.

5 variations to try with this prompt

Swap the subject — replace any [PRODUCT] or [SUBJECT] placeholder with your specific item. Be concrete: "matte navy ceramic mug" beats "a mug".
Change lighting language — try "golden hour rim lighting" or "harsh overhead studio fluorescent" instead of "soft studio lighting" to dramatically change mood.
Adjust camera — swap "eye-level close-up" to "overhead 90° flat-lay" or "low-angle hero shot" for different commercial framings.
Try a different visual register — append "rendered as a high-fashion magazine editorial" or "rendered as a 90s film photo with grain" to shift aesthetic without changing composition.
Change aspect ratio — switch `size: 1024x1024` to `1024x1536` (vertical) for IG Story / RedBook covers, or `1536x1024` (horizontal) for blog headers.

When to use this prompt (and when not to)

Use this prompt for: Xiaohongshu cover, Instagram square / 4:5, Pinterest pin, story tile. It's specifically tuned for those surfaces — using it elsewhere (e.g., a children's book) will produce technically correct but contextually wrong output. Skip it when you need: hands-on UGC photography (use a real camera), or content where text accuracy beats visual polish (gpt-image-2 still occasionally drops a letter on long strings). Pair it with: a follow-up edit on /edit to swap variants, or with the visual Prompt Lab on /lab to rebuild from atoms.

Common mistakes to avoid

Verbose adjective stacking — "beautiful, stunning, gorgeous, professional" all degrade output. The model interprets each as a separate weight competing for attention. One precise adjective beats five generic ones.
Over-specifying impossible camera — "8K hyperrealistic, fish-eye, telephoto" is contradictory; the model picks one and ignores the rest, often unpredictably. Pick one camera language at a time.
Quoting text without isolation — for prompts with rendered text, wrap the text in explicit quotes and place it in its own clause. `the text "GOOD MORNING" rendered cleanly in the lower third` beats `says good morning at bottom`.
Western aesthetic on Chinese platforms — Xiaohongshu / RedBook covers have specific conventions (large-character title overlay, soft-focus background, vertical 3:4). Don't apply Instagram conventions here.

FAQ

Is gpt-image-2 free to use through this site?

Yes — anonymous users get 1 generation per browser per day. Sign in with GitHub for 2/day, ⭐ star the open-source repo for 3/day. No credit card needed at any free tier.

How long does each generation take?

30–60 seconds for fresh generations on the 1024x1536 canvas. Cached prompts (same prompt + same size) return instantly and don't consume your quota.

Can I edit the generated image?

Yes — drop the result on /edit with a follow-up instruction (e.g., "swap background to night-sky"), or use /inpaint with a mask for surgical edits.

What's the prompt license?

All prompts in awesome-gpt-image-2-playground are CC-BY-4.0. Use them in any product, commercial or otherwise, with attribution.

Is the source code open?

Yes — repo on GitHub is MIT-licensed for code, CC-BY-4.0 for prompt content. Star the repo to support and to unlock 3 generations/day.

中文版 · ZH-CN

为什么这个 GPT Image 2 提示词有效

这个提示词是为小红书、Instagram、TikTok、Pinterest 上的创作者设计的。每个词都在做三件事中的一件:定义主体、控制光线/相机,或约束风格 — 没有冗余,核心目标是抗住 0.3 秒滑动停留测试,赢得收藏 / 二次传播。在 gpt-image-2 上稳定生效的关键是:模型对名词短语的解析比形容词堆砌更可靠。例如"放置在纯白背景中央(centered on pure white seamless background)"这种结构化表述,比"看起来干净"这种模糊形容更稳。中文小红书封面规则与西方 Instagram 美学差异巨大 — 必须显式声明纵横比与字体选择。

5 个值得尝试的变体

替换主体 — 将提示词中的 [PRODUCT] 或 [SUBJECT] 占位符替换为你的具体物品。"哑光海军蓝陶瓷杯"比"一个杯子"效果好十倍。
改变光线表达 — 把"柔和演播室光"换成"金时刻轮廓光"或"刺眼顶部荧光",氛围会发生剧烈变化。
调整相机 — 从"平视特写"切到"俯拍 90° 平面布局"或"低角度英雄镜头",得到完全不同的商业构图。
切换视觉调性 — 在末尾加上"以高时尚杂志编辑风格呈现"或"以 90 年代胶片颗粒感呈现",可以在不改构图的情况下换风格。
改变纵横比 — 把 `size: 1024x1024` 切到 `1024x1536`(竖版,适合 IG Story / 小红书封面)或 `1536x1024`(横版,适合博客 banner)。

什么场景下用(以及不用)

适合用于:小红书封面、Instagram 方图 / 4:5、Pinterest 钉图、Story 拼图。这个提示词专门针对这些场景调过 — 用在不匹配的语境(比如儿童绘本)会产出"技术上对、语境上错"的图。不适合的场景:UGC 真实手持摄影(用真相机更合适),或者文字精确度优先于视觉品质的内容(gpt-image-2 在长字符串偶尔会丢一个字母)。配合使用:在 /edit 上做 follow-up 编辑切换变体,或在 /lab 用原子重组重写整个提示词。

常见误区

形容词堆砌 — "美丽、惊艳、华丽、专业"叠在一起会降低输出质量。模型把每个都当独立权重在分配注意力,一个精准的形容词胜过五个通用的。
矛盾相机参数 — "8K 超写实、鱼眼、长焦"自相矛盾,模型会随机挑一个忽略其他。一次只用一种相机语言。
文字未隔离引用 — 需要渲染文字时,用引号包裹并独立成句。`the text "早安" rendered cleanly in the lower third` 远胜于 `底部说早安`。
直接套用西方平台美学 — 小红书封面规范(大字标题压图、虚化背景、3:4 竖版)和 Instagram 美学差异巨大,要显式声明,不要默认套 IG 风格。