Automatic subtitles that hear every word.

Auto-detect the language, transcribe with word-level precision, and keep captions synced to your edit. Automation for 95%, an editor for the rest.

How it works

01

Auto-detect or choose

Leave the language on auto-detect, or pick from 164 recognized languages when you want to be explicit.

02

Generate

The whole timeline's audio is transcribed in one pass; every word individually timed.

03

Review the transcript

The transcript panel shows every line against the playhead. The weak spots of any engine (names, jargon) are one double-click to fix.

Why word-level timing matters

Line-timed captions can only fade in and out. Word-timed captions can highlight each word as it's spoken (the style that carries short-form video) and they survive editing: split a caption and each word keeps its own clock.

Questions

How accurate is automatic transcription?
Very strong on clear speech; accents and crosstalk lower any engine's odds. Aulyr pairs automation with a real editor precisely because the last 5% is yours.
Which languages are detected?
164 source languages, auto-detected or set explicitly.
Does it add punctuation?
Yes: transcription includes punctuation and capitalization.
What if my video has no speech?
Aulyr tells you it found no speech instead of inventing captions.