agentsclimarketplace
Sign in

Video content extractor

Skill LiHongwei-cn/lihongwei-cn/mundo-cloud/skills/uncategorized/video-content-extractor

Extract key frames from MP4 videos at configurable intervals, run Tesseract OCR, and generate structured Markdown reports with video metadata and timestamped text transcripts.From its SKILL.md

Install
npx -y skills add LiHongwei-cn/lihongwei-cn --skill video-content-extractor

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 3 stars3 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

4.2 KB, 765 tokens by cl100k_base, as published. Nobody here has run it

Video Content Extractor

Overview

Automatically extracts key frames from MP4 video files at configurable time intervals, performs OCR text recognition on each frame, and generates a structured Markdown report. The report includes video metadata (duration, resolution, codecs) and frame-by-frame OCR transcripts with timestamp references.

This skill is designed for Codex CLI and requires FFmpeg and Tesseract OCR installed on the local machine.

When to Use This Skill

  • Use when you need to extract text content from video presentations, lectures, or screencasts.
  • Use when you want to create searchable transcripts from video files without embedded subtitles.
  • Use when you need to analyze video content programmatically and generate structured summaries.
  • Use when the user asks to "read what is on screen" or "extract the content from this video."

How It Works

Step 1: Analyze Video Metadata

The skill uses ffprobe to extract video metadata: duration, resolution, frame rate, codec information, and file size.

Step 2: Extract Key Frames

Using FFmpeg, the skill captures frames at the configured interval (default: every 30 seconds). Each frame is saved as a timestamped JPEG image.

Step 3: OCR Text Recognition

Each extracted frame is processed by Tesseract OCR. If the default PSM mode returns no meaningful text, it falls back to fully automatic page segmentation.

Step 4: Generate Markdown Report

All extracted data is assembled into a structured Markdown document.

Examples

Example 1: Basic Extraction

Agent prompt: Use the video-content-extractor skill to extract content from lecture.mp4

Output generates lecture.md and lecture_frames/ directory.

Example 2: Custom Interval

Parameters: video_path, output_dir, interval(seconds), lang Extract every 60 seconds with English-only OCR: python scripts/extract_video.py recording.mp4 ./output 60 eng

Example 3: Bilingual Content

Extract with default Chinese + English OCR: python scripts/extract_video.py lecture.mp4 . 15 chi_sim+eng

Best Practices

  • Use shorter intervals (10-15s) for fast-paced content with frequent text changes.
  • Use longer intervals (30-60s) for presentation slides or slow lectures to reduce duplicate frames.
  • For Chinese content, ensure Tesseract Chinese language pack is installed (chi_sim).

Limitations

  • Requires FFmpeg and Tesseract OCR to be installed and accessible via PATH.
  • Tesseract OCR accuracy depends on video quality, text size, and font clarity.
  • Does not extract audio or perform speech-to-text transcription.
  • Frame extraction is time-based (not scene-change-based), which may produce near-duplicate frames.
  • Large videos with short intervals can generate many frames - ensure sufficient disk space.

Security and Safety Notes

  • This skill only reads video files and writes extracted frames and Markdown reports.
  • It does NOT send any data over the network - all processing is local.
  • FFmpeg and Tesseract are invoked with fixed, pre-vetted arguments.
  • The skill does not modify or delete the original video file.

Common Pitfalls

  • Problem: Tesseract returns garbled text Solution: Ensure the correct language pack is installed. Run tesseract --list-langs to verify.

  • Problem: FFmpeg fails with "not found" Solution: Make sure FFmpeg is on PATH. Run ffmpeg -version to verify.

  • Problem: OCR is slow on large videos Solution: Increase the interval parameter to reduce frames processed.

Related Skills

  • @media-summarizer - For summarizing video content using visual and audio cues.
  • @document-ocr - For OCR on static images or scanned documents without video processing.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Gives 0 of the 12 instructions most video audio skills give in 765 tokens

Counted across 619 of the 725 authors here whose files we hold, read 2026-09-06

  • Read product marketing context firstin 13 of 619, across 7 files
  • Define the core visual thesis in one sentencein 11 of 619, across 3 files
  • Break the concept into 3 to 6 scenesin 11 of 619, across 3 files
  • Render the smallest working version firstin 11 of 619, across 3 files
  • Start with a low-quality smoke test renderin 11 of 619, across 3 files
  • Add captions for accessibility and engagementin 11 of 619, across 5 files
  • Write the scene outline before writing codein 11 of 619, across 3 files
  • Specify subject, action, camera, style, and moodin 11 of 619, across 5 files
  • Decide what each scene provesin 10 of 619, across 2 files
  • Export one clean thumbnail framein 10 of 619, across 2 files
  • Pick the right tool for the jobin 10 of 619, across 4 files
  • Run the test suite before proposing a fixin 8 of 619, across 7 files

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.