Click a word, hear the moment
Click to skip is on by default: click any word and the audio or video jumps there. Each paragraph also has its own play button.
Reviewed by David Dat Nguyen, CXO Last updated
Every word knows
when it was said.
Sonix stores a start and end time for every word, to the hundredth of a second, so you can click any word to hear it, cite the exact moment and export timecodes wherever you need them.
Paragraph timecodes by default, interval timecodes from five seconds to ten minutes, frames for video work, and the same times in the API.
30 free minutes on every new account, no credit card.See pricing
Three things word-level timing gives you.
Click to skip is on by default: click any word and the audio or video jumps there. Each paragraph also has its own play button.
A quote carries its time in the transcript and in the export, so a fact-check, a court reference or an edit note points at the moment, not the page.
A timecode at every paragraph, every 30 seconds, before each speaker, as hundredths or frames, or per word in CSV and the API.
How three kinds of teams use the times.
Quotes that can be checked against the recording.
Finding the take by its words.
A working transcript where every statement has a time.
The settings and limits, as they are in the product today.
Export the transcript to Premiere XML or Final Cut Pro X so every word sits on the timeline, then find a take by searching its words. Set the timecode display to frames at your project's frame rate, and use the start-timecode offset if the recording's clock did not begin at zero.
Each word's start and end are stored to the hundredth of a second, and that is what the editor shows, the exports print and the API returns. Timing follows the audio, so a word spoken over another voice is placed where the engine heard it begin.
Yes. In DOCX and TXT exports choose a timecode interval from every 5 seconds to every 10 minutes, with or without the paragraph timecodes, and optionally before each speaker name.
Each paragraph's timecode appears in the DOCX and PDF, and the editor shows the time of any word you click, so a reference can name the minute and second. For an exact span, the CSV export lists the start and end of every word.
Words you leave alone keep their times. After heavy editing, Tools has a realign action that regenerates every word's timecode against the audio; it uses transcription time for the length of the file.
The feature this page belongs to, and its neighbors.
Upload a recording and jump through it by its words. 30 free minutes, no credit card.
AI transcription and translation in 54+ languages.