Editing an utterance as a line of text
Utterances can now be edited as continuous text, with audio alignment and annotations retained for unchanged words.
Work done September 2–5, 2026 · Last edited September 25, 2026
On this page
Arc now supports editing each utterance as a continuous line of text. Words can be corrected, added, and removed within the transcript, bringing familiar text editing to the work of checking an automatic draft against the recording. Individual words can still be marked as fillers, retraces, or errors.
Keeping words connected to the recording
When an edit is saved, Arc matches the revised text to the existing words. Words that have not changed retain their audio timing, markings, and phonetic transcriptions.
Speech identified in the recording but missing from the draft appears as a question mark in the text. Replacing it with the words heard connects that transcription to the corresponding audio. Unintelligible speech can be entered as xxx.
Annotations across the transcript
Annotation layers can now be displayed across all utterances, including those not currently being edited. This makes it possible to read through the transcript with the relevant annotations in view.
