Video to Context

Video to Context is a local command-line tool for turning screen recordings, voiceovers, and voice memos into reusable context packages. It extracts local audio with FFmpeg, transcribes it with whisper.cpp, and writes a small bundle that can be read by a person or handed to an AI system for later analysis.

• Source: byronwall/video-to-context
• Voice memo preset: v2c --voice-memos

How it works

The CLI accepts a single media file or a folder of recordings. For videos, it can extract screenshots, create a contact sheet, transcribe the audio track, and produce an HTML report that interleaves visuals and narration. For audio-only files, it skips the visual pipeline and focuses on the transcript, timeline, and source lineage.

Voice memos get a dedicated preset:

v2c --voice-memos

That flag auto-detects the likely Apple Voice Memos folder, writes to ~/.v2c-voice-memos, avoids copying private source audio by default, skips screenshot work, opens the finished report, and uses a manifest so identical reruns do not transcribe the same files again.

Highlights

• Local-first processing with FFmpeg and whisper.cpp
• Supports both screen recordings and audio-only memos
• Directory mode combines many files into one timeline with source lineage
• Idempotent output folder with .v2c-manifest.json
• One-command voice memo workflow through --voice-memos
• HTML and markdown outputs designed for later human or AI review

Why I use it

Voice memos are a fast way to capture design notes, implementation thoughts, and narration that would otherwise stay trapped in an app. Screen-recording voiceovers are similarly useful, but only after the audio becomes searchable. Video to Context turns those recordings into a durable text artifact without uploading them or splitting the workflow across separate tools.