Skip to main content
Gladia provides word-level timestamps by default in pre-recorded transcriptions. Under each utterance, you’ll find a words property like this: