wjs-reframing-video
Reframe videos between landscape and portrait by cropping a dynamic band that follows whoever is speaking, detected via mouth motion analysis. Ideal for repurposing interviews and podcasts across short-form platforms without rotating the source.
wjs-reframing-video converts videos between horizontal and vertical by cropping and tracking the active speaker.
AI-generated summary based on this skill's SKILL.md
Decision gist · record as of 2026-07-23
wjs-reframing-video converts videos between horizontal and vertical by cropping and tracking the active speaker. Reframe videos between landscape and portrait by cropping a dynamic band that follows whoever is speaking, detected via mouth motion analysis. Ideal for repurposing interviews and podcasts across short-form platforms without rotating the source.
Use it when
- Yes.
- wjs-reframing-video uses mouth motion analysis to detect who is speaking in multi-person videos.
Verify before relying
Read SKILL.md below before installing (4 files). Open directory: indexed for reading, not audited.
Install
jianshuo/claude-skills/wjs-reframing-video · repository language: Python
Open directory. Skills are indexed for reading, not audited. Review a skill's body before installing it.
Frequently asked questions
AI-generated answers based on this skill's SKILL.md and metadata
How does wjs-reframing-video convert 16:9 video to 9:16 portrait?
wjs-reframing-video reframes landscape footage into portrait by cropping a dynamic vertical band that intelligently follows the active speaker. Instead of letterboxing or rotating the source, it analyzes mouth motion to detect who is speaking and continuously adjusts the crop window to keep the talking head centered, making it ideal for repurposing interviews and podcasts for short-form platforms.
Can wjs-reframing-video make horizontal video vertical for TikTok?
Yes. wjs-reframing-video specializes in repurposing horizontal video content for vertical short-form platforms like TikTok and Instagram Reels. It automatically crops and reframes your landscape footage into portrait aspect ratio while tracking the active speaker, so your content stays engaging without awkward letterboxing or manual cropping.
How does wjs-reframing-video auto-crop video tracking the active speaker?
wjs-reframing-video uses mouth motion analysis to detect who is speaking in multi-person videos, then automatically crops a dynamic vertical band that follows that person's face. This speaker-detection approach keeps the talking head centered throughout the reframe, making it perfect for interviews, podcasts, and panel discussions converted to vertical format.
What aspect ratios does wjs-reframing-video support for reframing?
wjs-reframing-video handles conversion between landscape and portrait orientations, including 16:9 to 9:16 and other standard aspect ratios. It reframes interview and podcast footage to match your target aspect ratio without letterboxing, using intelligent speaker tracking to preserve the most important content in the crop.
Can wjs-reframing-video generate metadata for video reframing workflows?
wjs-reframing-video can generate crop plans and sidecar metadata to support your video reframing workflows, enabling integration into larger production pipelines. This metadata output allows you to document and automate the reframing decisions for batch processing or downstream tools.
What license does wjs-reframing-video use?
wjs-reframing-video is released under the MIT license, making it free to use, modify, and distribute for both commercial and personal projects with minimal restrictions.
SKILL.md
Rendered from the published skill. Quoted content, verbatim.
wjs-reframing-video
Convert a video's orientation by cropping a narrow band from the source — not by physically rotating it. The crop window follows the active speaker (the face whose mouth is moving), not just the largest or most-confident face. A .crop.json sidecar records the crop plan, the per-segment speaker decisions, and the parameters used. The original input is never modified.
When to use
- Repurposing a 16:9 podcast / interview / talk for vertical short-video platforms (WeChat Channels 视频号, Douyin 抖音, Xiaohongshu 小红书, YouTube Shorts, TikTok,
(truncated - see the full file via the links below)
File tree — 4 files
wjs-reframing-video/SKILL.md
wjs-reframing-video/models/blaze_face_short_range.tflite
wjs-reframing-video/models/face_landmarker.task
wjs-reframing-video/scripts/crop.py
Let your AI agent find skills like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 56,283 agent skills by what they can do, searchable in plain language.
wish › “Convert video between horizontal and vertical orientations while tracking the active speaker”
Give your agent the search over MCP, or paste the wish link into any chat. No install? Search from any chat →
Related skills
Video Action Recognition provides a pipeline for identifying human actions in video by extracting body keypoints and applying rule-based classifiers. It uses MediaPipe Pose to detect 33 body landmarks, then analyzes movement patterns to recognize actions like arm flapping, head banging, and spinning without requiring any model training.
Segment long-form video into multiple stand-alone short clips by reading the SRT transcript and identifying topic boundaries. The skill cuts and crops with accurate-seek precision, extracting raw clips plus per-clip subtitles as a hand-off package for downstream post-production tools like overlaying and styling.
Clipify turns long-form videos into vertical social clips by identifying clip-worthy moments via transcription, automatically reframing 16:9 footage to 9:16 with motion-based face tracking, and applying word-level captions. It detects punchlines, reversals, and audio peaks to propose 3–5 candidates, then renders with hardware acceleration on macOS.
Compose post-production overlays—captions, covers, illustrations, and calls-to-action—directly onto video clips using HyperFrames, rendering everything in one pass to preserve quality. Works seamlessly downstream of segmentation workflows, turning raw clips into upload-ready MP4s with HTML/CSS captions, AI-generated covers, and timed motion graphics.
Merge multiple camera angles into a single rendered video using audio energy to drive real-time cam selection. The skill reads sync metadata from each source, analyzes per-second loudness, and produces hard cuts or hard cuts with a corner PiP inset. Requires synced footage with `.sync.json` sidecars from wjs-syncing-multicam.
This skill converts a written article into a polished short-form video optimized for vertical platforms. It combines Volcano TTS Chinese narration, HyperFrames-powered scene animations, and AI-generated abstract watercolor backgrounds to produce 1080×1920 MP4 files ready for YouTube Shorts, TikTok, and similar services. The output emphasizes visual contrast and pacing over template repetition, enforcing diverse layouts and typography across 5–10 scenes.
More skills wjs-localizing-video (MIT) · aliyun-vidu-video (MIT) · Ffmpeg Patterns (unlicensed) · Media Transcription (unlicensed) · wjs-syncing-multicam (MIT)