hindi-video-transcriptlisted

# Hindi Video Transcript Turn a Hindi/Hinglish YouTube video or playlist into two files per video: an exact Devanagari transcript and a Roman Hindi one. ## Do not use when - The value is on screen — silent demos, visual walkthroughs - There is no caption track - The video is not Hindi — fetch its track directly, skip transliteration ## Layout One base folder holds one shared `tools/` and one folder per subject. Glossaries are per subject and must never merge, so each keeps its own. ``` <base>/ ├── tools/ copied once, every project uses it │ translit.py │ convert.py ├── Trading Course/ │ ├── 01 - Charting.txt what the user reads │ ├── glossary.py this subject's own │ ├── _glossary-todo.txt delete when clean │ └── raw/ never opened by hand │ 01 - Charting (hi-orig).txt │ 01 - Charting.json3 └── Cooking Course/ new subject = new folder, nothing else to set up ``` ## Step 1 — Ask first One `AskUserQuestion` call, before downloading anything: | Ask | Note | |---|---| | YouTube link | Detect single video or playlist | | Base folder | Full path. Never guess, never default | | Timestamps in the Roman file? | Default **no** | | How many videos | Playlist only: one / all / batches of N | Then list `<base>` and ask once more: an existing subject folder, or a new one and its name. Name it after the content, never after this skill. ## Step 2 — Tooling Note the OS first. Every command