Reframe Video to Vertical or Square

Reframe local video with a center or manual focal crop, or black padding. Export bounded even-sized video and a before/after geometry manifest.

ClaudeHermes AgentCodexCursorMIT (modify, redistribute and use commercially)Python

What it does

Convert local landscape video to approximate 9:16 vertical or 1:1 square framing with a fixed center/manual focal crop, or black padding that retains the whole image. Optionally set an even output height. A geometry manifest includes source dimensions, crop, planned dimensions, command and measured output dimensions.

How to use

Choose crop or pad and vertical or square. Focal x/y fractions range from 0 to 1 and affect crop only. Review a dry run before exporting to a new directory.

Requirements

Python 3.10+, FFmpeg and ffprobe installed separately. No Python packages, uploads or model downloads. FFmpeg must include H.264 and AAC encoding.

Caveats

No AI face detection or tracking; the focal position is fixed throughout. Dimensions are even and capped at 4096 per side; rounding slightly changes aspect ratio. Rotated metadata and non-square pixels are rejected until normalized. Only the first video and optional first audio are retained; captions, global container tags and chapters are dropped. Does not overwrite existing directories or download arbitrary YouTube URLs.

How to install

  1. Unzip the downloaded file (it contains a video-vertical-reframe/ folder with SKILL.md).
  2. Copy the whole folder into your agent's skills folder.
    • Claude: ~/.claude/skills/ (per project: .claude/skills/)
    • Hermes Agent: ~/.hermes/skills/
    • Codex: ~/.agents/skills/ / Cursor: ~/.cursor/skills/
    • Claude app (claude.ai): upload the ZIP in Customize → Skills
  3. Restart the agent and check that the skill appears in its skill list.

Step-by-step install for Windows / Mac and each agent

Seller

動画編集が仕事

動画編集の準備から仕上がりの確認まで、毎日の作業を支援するスキルを制作しています。切り出し・字幕づくり・音声調整・文字起こしを、元の素材を残しながら進められるようにします。

10 published skills

Seller profile and all skills