video-understand

You can transcribe video audio offline without an API key by using a local Whisper model to process the audio track directly on your machine. This Skill extracts audio and transcribes it entirely locally without relying on cloud-based APIs.

Extract keyframes from a video using ffmpeg by applying frame extraction at scene changes. This Skill automates ffmpeg frame extraction, supporting multiple extraction modes to isolate keyframes for visual analysis and scene inspection.

You need openai-whisper installed only if you want to transcribe audio locally. The Whisper model is optional; frame extraction and scene inspection features work independently through ffmpeg without requiring openai-whisper.

The best way to perform offline transcription for educational videos is using a local Whisper model to process the audio without cloud services. This Skill handles offline transcription specifically for education, research, or content production contexts.

You can extract frames at scene changes without cloud-based APIs by running ffmpeg locally to process the video file. This Skill operates fully offline, applying ffmpeg frame extraction to detect scene changes and isolate keyframes.

Offline video transcription with Whisper requires ffmpeg, ffprobe, and an optional Whisper model installed locally. Processing happens entirely on your machine, so performance depends on your local hardware capabilities rather than cloud resources.