SuperAPI AI Directory
Partner中文AI newsFavorites
Skillsarticle-cover-designer

article-cover-designer

Design & creation

View repository

Updated 10-16 15:13

About this skill

Article Cover Designer (Article and Video Cover Designer)

Minimalist and highly efficient conversion of articles or video subtitles into cover image generation prompts with high aesthetics, high conversion, and high eye-catching appeal.

Core Design Principles

Completely abandon lengthy and complex theoretical statements, retaining only what image generation models (such as GPT Image 2/2.5, Midjourney, etc.) need mostCore Business and Design Fields：

  1. Core Summary: extract about 200 words of the core argument, main storyline, or background logic;
  2. Audience Inference: infer from the content the precise target audience；
  3. Business Scenarios：
    • Article Cover: default to 公众号文章封面(preferred 2.35:1 or 21:9）、X 文章封面（5:2) or 小红书封面（3:4）；
    • Video Cover: when the user explicitly requestsvideo cover/thumbnail, or inputssubtitle file（.srt、.vtt, subtitle verbatim transcript/timeline text), automatically identify as 视频封面(horizontal screen recommended 16:9, vertical short video recommended 9:16 or 3:4）；
  4. Style and Color Scheme Injection: from the entire library of 331 hand-drawn styles（#001–#331) and 36 classic color schemes（C-01–C-36) match or have AI recommend;
  5. Fixed Design Instructions：
    • Ordinary Article Cover: inject “design the metaphor first before generating the image, the main title is obvious, small text is minimal or absent. You design the rest for me.” to let the model natively and autonomously compose and lay out;
    • Video Cover / Subtitle Input: must inject into the prompt “video cover design, strong internet feel, eye-catching. Design the metaphor first before generating the image, the main title is obvious, small text is minimal or absent. You design the rest for me.” to strengthen visual impact and click-catching appeal.

Workflow and Operation Steps

Input Recognition and Scenario Determination

  • Article Input: the user provides article text, Markdown content, or a local article path (such as D:\path\to\article.md) -> enable [Article Cover Mode];
  • Video / Subtitle Input: the user explicitly states it is a “video cover/thumbnail/Bilibili cover/YouTube cover/short video cover,” or provides a subtitle file (such as .srt, .vtt, .txt subtitles, containing a timeline verbatim transcript of dialogue) -> automatically determine as [Video Cover Mode]；
  • (optional) specify style number (001–331) and theme color number (C-01–C-36). If not specified, AI intelligently recommends the optimal combination based on the content.

Execution Process

  1. Extract a 200-word Core Summary: Summarize the core event/core viewpoint, causal logic, and cognitive insights (if subtitles, extract the video's core theme and highlights);
  2. Infer audience profile: Based on knowledge depth and needs, output a 1–2 sentence precise audience positioning;
  3. Determine business scenario and canvas size：
    • Official account article cover: recommended canvas size 2.35:1 or 21:9；
    • X article cover: recommended canvas size 5:2；
    • Xiaohongshu image-text cover: recommended canvas size 3:4；
    • Video cover (landscape, e.g. Bilibili/YouTube/WeChat Channels landscape): recommended canvas size 16:9；
    • Video cover (portrait, e.g. Douyin/Kuaishou/Xiaohongshu/WeChat Channels portrait): recommended canvas size 9:16 or 3:4；
  4. Inject style and color: Output the recommended specific style (e.g. #018 · Minimal Deadpan Dialogue Cartoon) and theme color (e.g. C-01 经典蓝) along with the reasons for the recommendation;
  5. Assemble the standard prompt: Output the Chinese and English prompts directly according to the prescribed template (if it is a video/subtitle scenario, strictly include "Video cover design, strong internet feel, eye-catching"）。

Standard prompt output template

Mode A: Article cover (default)

Chinese prompt:

文章摘要：{200字左右的核心摘要}。
受众：{推算出的核心受众}
业务场景：{公众号文章封面 / X 文章封面 (5:2) / 小红书封面 (3:4)}。

风格：{风格编号及名称，如 #018 Minimal Deadpan Dialogue Cartoon}
主题色：{主题色编号及名称，如 C-01 经典蓝}
先设计隐喻再出图，主标题明显，小字少或者没有小字。 其他你帮我设计。

请设计文章封面

English Prompt：

Article Summary: {200-word concise summary in English}.
Audience: {Inferred target audience in English}.
Scenario: {WeChat Official Account Cover / X Article Cover (5:2) / Xiaohongshu Cover (3:4)}.

Style: {Style ID and Name}
Theme Color: {Theme Color ID and Name}
Design a visual metaphor first, then generate the image. Ensure the main title is bold and prominent, with few or no small text. Pick the rest of the design for me.

Please design the article cover.

Mode B: Video cover / subtitle input (triggered by video cover or subtitle file)

Effective conditions：

  1. The user explicitly states that a "video cover," "video thumbnail," "Bilibili cover," "YouTube cover," or "short video cover" is needed;
  2. Or the user input is a subtitle file (.srt、.vtt, verbatim transcript subtitles containing a timeline).

Mandatory prompt requirements: The prompt must state:Video cover design, strong internet feel, eye-catching。

Chinese prompt:

视频摘要：{根据视频字幕或内容提炼的200字核心摘要}。
受众：{推算出的核心受众}
业务场景：{视频封面 (16:9) / 竖屏短视频封面 (9:16 或 3:4)}。

风格：{风格编号及名称，如 #018 Minimal Deadpan Dialogue Cartoon}
主题色：{主题色编号及名称，如 C-01 经典蓝}
视频封面设计，网感强，吸引人眼球。先设计隐喻再出图，主标题明显，小字少或者没有小字。 其他你帮我设计。

请设计视频封面

English Prompt：

Video Summary: {200-word concise summary extracted from subtitles/video content in English}.
Audience: {Inferred target audience in English}.
Scenario: {Video Cover / Thumbnail (16:9) / Short Video Cover (9:16 or 3:4)}.

Style: {Style ID and Name}
Theme Color: {Theme Color ID and Name}
Video cover design, high clickability and viral internet appeal, visually captivating and eye-catching. Design a visual metaphor first, then generate the image. Ensure the main title is bold and prominent, with few or no small text. Pick the rest of the design for me.

Please design the video cover.

Delivery and follow-up action guidance (CTA)

After outputting the prompt, proactively provide a minimalist dual-track guide:

---

💡 **封面设计方案已完成！您可以选择：**
1. **【方式 A · 自主生图】**：复制上方提示词，前往您常用的生图工具出图；
2. **【方式 B · 全自动生图】**：直接对我说 **“全自动出图”**，我将调用生图工具为您全自动生成封面图片！