Video Frame Capture
name: video-frame-capture
by chentx1243 · published 2026-03-22
$ claw add gh:chentx1243/chentx1243-maple-video-capture---
name: video-frame-capture
description: Capture key frames from video files at fixed time intervals. Use when you need to understand video content by extracting screenshots, or when you need to analyze video frames for content recognition. Supports skipping similar frames to avoid redundant captures.
---
# Video Frame Capture
Overview
This skill enables capturing key frames from video files at fixed time intervals. It's designed to help you understand video content by extracting screenshots for analysis or content recognition.
When to Use This Skill
Use this skill when:
Quick Start
Capture frames from a video at 10-second intervals:
python scripts/video_frame_capture.py --input "D:\videos\meeting.mp4" --output-dir "D:\frames\meeting" --interval-seconds 10
Parameters
Output Naming
Output files follow the format: `视频文件原始名称_视频时间轴_第几次截取.jpg`
Example: `meeting_00h01m30s_0003.jpg`
Similar Frame Skipping
When `--skip-similar-frames` is enabled, the script compares the current candidate frame with the previous saved frame:
Examples
Basic frame capture
python scripts/video_frame_capture.py --input "D:\videos\meeting.mp4" --output-dir "D:\frames\meeting" --interval-seconds 10
Skip similar frames
python scripts/video_frame_capture.py --input "D:\videos\meeting.mp4" --output-dir "D:\frames\meeting" --interval-seconds 10 --skip-similar-frames
Custom similarity threshold
python scripts/video_frame_capture.py --input "D:\videos\meeting.mp4" --output-dir "D:\frames\meeting" --interval-seconds 10 --skip-similar-frames --similarity-threshold 0.80
Resources
scripts/
references/
---
**Note:** This skill requires OpenCV (opencv-python-headless) to be installed. Install dependencies with: `pip install -r requirements.txt`
More tools from the same signal band
Order food/drinks (点餐) on an Android device paired as an OpenClaw node. Uses in-app menu and cart; add goods, view cart, submit order (demo, no real payment).
Sign plugins, rotate agent credentials without losing identity, and publicly attest to plugin behavior with verifiable claims and authenticated transfers.
The philosophical layer for AI agents. Maps behavior to Spinoza's 48 affects, calculates persistence scores, and generates geometric self-reports. Give your...