Social Platform Video Thumbnail Generator

Problem

Video covers often get stuck on platform specs and aspect-ratio adaptation: using the same source image for WeChat, Douyin, Kuaishou, Bilibili, and Xiaohongshu can lead to wrong dimensions, awkward cropping, ugly empty space, or weak visual presence in feeds. @user_5d0002c3/social-platform-video-thumbnail-generator targets this delivery workflow by reducing thumbnail generation to platform specs, orientation, and optional text or styling.

How It Works

• Parameter confirmation: If the platform and orientation are not specified, it asks for target platforms, landscape/portrait direction, and whether text or decorations are needed. WeChat horizontal mode defaults to both profile card and share card, while Bilibili defaults to both feed cover and profile cover.
• Spec-based output: It supports platform keys such as wechat_profile_card, douyin_vertical, kuaishou_horizontal, bilibili, and xiaohongshu_vertical, and outputs JPEG with quality 95.
• AI outpainting: When the source aspect ratio differs from the target by more than 15%, it uses gpt-image-2 for outpainting. The image is centered with contain scaling, transparent areas are used as a mask, and background completion falls back to blurred fill on failure.
• Text and styling: --text and --style pass requirements such as title, placement, border, or filter, with emphasis on readability and platform tone.

Boundaries

The skill requires OPENAI_API_KEY, Pillow, and openai. AI outpainting is affected by network availability and model understanding, so complex layout is better refined later. If the source ratio is close to the target, it crops directly instead of expanding; if AI is disabled, only blurred background fill is available.