メニュー
Extract text from video files by sampling frames and running Apple Vision OCR, with optional perceptual deduplication
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Use when the user asks to "browse a website", "go to a URL", "fill out a form", "take a screenshot", "click on something", "extract data from a page", "automate a browser task", "control the desktop", "use the computer", or any task involving web pages or desktop applications.
Use when the user wants to make a phone call, leave a voicemail, call a business, reach someone by phone, or navigate a phone menu. Triggered by "/call" followed by a phone number and purpose.
Start or continue a progressive learning session. Use when the user says "learn", "let's learn", "continue learning", "let's continue", "pick up where we left off", "teach me", "next lesson", or wants to resume their learning curriculum. Requires a phase map to exist in docs/phases/.
Design progressive learning curricula and generate project phase maps. Use when the user says "learn-design", "design curriculum", "plan learning path", "set up next project", "what should I learn next", or wants to start learning a new technical area. Also triggers between projects to generate the next project's phase map calibrated to the learner's profile.
This skill converts any content source into podcast episodes. Triggers: sending a URL, PDF, YouTube link, book, or file with intent to listen; phrases like 'podcast this', 'make this a podcast', 'listen to this', 'generate a podcast'; sending content with instructions like 'split by chapters', 'make a series'. Handles single articles, multi-chapter books, YouTube playlists, and any document format.
This skill should be used when the user asks to "read a paper", "understand a paper", "analyze a research paper", "prepare for paper discussion", "work through a paper", "do a literature review", or provides PDF paths to research papers. Guides deep paper understanding using Keshav's three-pass method with AI as a patient professor. Supports multiple papers for literature review.
| name | video-ocr |
| description | Extract text from video files by sampling frames and running Apple Vision OCR, with optional perceptual deduplication |
Extracts frames from a video at a configurable FPS using ffmpeg, optionally deduplicates visually similar frames using perceptual hashing (aHash), then runs Apple Vision OCR on each kept frame in parallel.
Use this skill when the user wants to extract text from a video -- for example lecture recordings, tutorial screencasts, or presentation recordings where on-screen text changes over time.
python3 ${CLAUDE_PLUGIN_ROOT}/scripts/video_ocr.py <video_file> \
[-o output.jsonl] \
[--fps 3.0] \
[--workers 8] \
[--dedupe] [--dedupe-threshold 0.15] [--hash-size 8] \
[--frames-out <dir>] \
[--markdown <output.md>] [--markdown-images]
| Argument | Default | Description |
|---|---|---|
video (positional) | required | Input video file path |
-o, --output | ocr_output.jsonl | Output JSONL file path |
--fps | 3.0 | Frames per second to extract |
--workers | 8 | Number of parallel OCR workers |
--dedupe | false | Deduplicate visually similar frames |
--dedupe-threshold | 0.15 | Max visual difference ratio (0-1) to treat as similar |
--hash-size | 8 | Perceptual hash size (hash_size x hash_size bits) |
--frames-out | None | Directory to save kept frames as JPGs |
--markdown | None | Optional Markdown output path |
--markdown-images | false | Embed frame images in Markdown (requires --frames-out) |
--dedupe is enabled, frames are compared using average perceptual hash (aHash) and similar consecutive frames are droppedframe, time_sec, and text fields)The deduplication feature uses average perceptual hashing (aHash):
hash_size x hash_size grayscale--dedupe-threshold are dropped