Skip to content
Get started

allyhub words

A recording's words: heard on this computer when it has none (or --fresh), printed by word index with the sounds between them; a stretch by seconds, heard again with --fresh

Terminal window
allyhub words [file] [options]
Argument Type What it is
<file> string The recording: an audio or video file
Option Type Default What it does
--lang string The language spoken (en, es, fr, de, zh…); left out, the cloud finds it
--fresh boolean false Hear the whole file again and write its words file; with –from and –to, only that stretch, writing nothing
--from number Show the words of the file from this second on
--to number Show the words of the file up to this second
--search list of string Where these words or phrases are said (case and punctuation ignored; a phrase’s last word may be a prefix), instead of the transcript
--script string A script to compare with: a video’s script in one language (what its parts say) or a text file; marks words it does not have, {like this}, and lines said more than once
--all boolean false Also list what the thresholds leave out: every stretch between words above the noise floor, every word’s inside, and every script line

--dir <value>: the project’s folder; default the one around where it is run. --json: print the result as JSON.

With --json, the command prints an object with these fields:

Field Type What it is
file string
wordsFile
Terminal window
allyhub words talk/host.mp4
allyhub words host.mp4 --from 23 --to 25
allyhub words host.mp4 --from 23 --to 25 --fresh
allyhub words host.mp4 --search "you know,um"

0 when it succeeds; the others are listed in Exit codes.

footage: Cutting recorded footage: a talking head, an interview

words → words edits → align → frames cuts → audio recording → render

references: Studying reference videos, and a piece against them

reference → reference compare → reference set → frames → frames compare → words → audio → align

Was this page helpful?