prompt logo

prompt

Free

AI prompt for designing social media image cards

FreeFree tier
Inputs: text, imageOutputs: image
Type
Open Source
Company
op7418

About prompt

A specialized system prompt for AI to act as a Social Card Designer, generating production-ready social media image cards from articles, copy, screenshots, and photos. It outputs single-file HTML that can be rendered to PNG via Playwright, targeting Xiaohongshu (Rednote) carousels and WeChat Official Account cover pairs. The prompt enforces a strict production pipeline: style system (Editorial Magazine or Swiss Internationalism), surface format (1080×1440, 2100×900, or 1080×1080 px), narrative arc, asset availability, and theme preset—all chosen upfront. It uses 28 registered layouts (16 for Style E, 12 for Style S) and 10 locked theme presets, with a clear image sourcing hierarchy (user-provided > Unsplash > Pexels > Flickr CC).

Key Features

Two locked style systems: Style E (Editorial Magazine) for narrative and lifestyle, Style S (Swiss Internationalism) for product reviews and data
Surface-first canvas discipline with three fixed formats: Xiaohongshu 3:4 carousel (1080×1440), WeChat header (2100×900), WeChat share card (1080×1080)
28 registered layout codes (M01–M16 for Style E, S01–S12 for Style S) – no invented structures
10 locked theme presets: 6 for Style E (e.g., Ink Classic, Kraft Paper) and 4 anchor colors for Style S (e.g., IKB Klein Blue, Safety Orange)
Image-source hygiene prioritizing user-provided images, then Unsplash, Pexels, Flickr CC
Output as single-file HTML rendered to PNG via Playwright, no frontend build chain
Clear production pipeline: confirm style, surface, narrative arc, asset availability, and theme preset before writing markup

Pros & Cons

Pros
  • Provides well-defined style systems ensuring visual consistency across cards
  • Includes 28 pre-registered layouts for common content types (image-led, quote, grid, pipeline, etc.)
  • Locked theme presets eliminate design decision fatigue and enforce brand cohesion
  • Outputs single-file HTML that can be easily converted to PNG for direct social media upload
  • Clear image sourcing hierarchy reduces visual clutter and legal worries
  • Targets two major Chinese social platforms with precise aspect ratios and guidelines
Cons
  • Strictly tailored to Chinese social platforms (Xiaohongshu and WeChat), limiting global applicability
  • Requires a capable AI model that can follow detailed layout instructions accurately
  • No built-in rendering engine; requires Playwright or similar tool to convert HTML to PNG
  • Limited to the style and layout systems defined in the prompt; customization beyond presets is discouraged
  • Adherence to rules depends on the AI's ability to enforce all constraints consistently

Best For

Designing carousel cards for Xiaohongshu (Rednote) from lifestyle articles, travel notes, or film reviewsCreating WeChat Official Account cover pairs (header + square share card) for blog posts or tutorialsProducing product review cards with data-driven layouts (KPI Tower, H-Bar Chart) under Style SGenerating before/after comparison cards for design portfolios or recipe stepsMaking quotation cards for personal observations or reading notes using Style E layouts

FAQ

What platforms does this prompt target?
It targets two dominant Chinese social surfaces: Xiaohongshu (Rednote) carousels and WeChat Official Account cover pairs, with specific aspect ratios and safe margins.
What are the main style systems?
Style E (Editorial Magazine) for narrative, lifestyle, travel, and reading content; Style S (Swiss Internationalism) for product reviews, data, tutorials, and AI tools.
How many layouts are available?
Style E offers 16 registered layouts (M01–M16), Style S offers 12 (S01–S12). Every card container must be tagged with its layout code.
Can custom hex values be used for colors?
No. Theme presets are chosen from locked options: 6 for Style E (e.g., Ink Classic, Dune) and 4 anchor colors for Style S (e.g., Lemon Yellow, Safety Orange). No arbitrary hex values.
How are images sourced?
User-provided images are first priority. If needed, the AI sources from Unsplash, then Pexels, then Flickr CC, in that order.