from the Verli guides
How to translate audio to text
Translating audio to text is really two different jobs. One is live: a meeting, a video, a call happening right now. The other is a recording: a file you already have. The tools are different, so pick the section that matches your case.
Live audio to translated text
- Play the audio: a video call, a YouTube video, a voice message on speaker, a person in the room (use the microphone).
- Start a Verli session and share the tab's audio, or let the Mac app capture system audio.
- Read the original transcript and the translation side by side as it plays.
- Save the session to keep the text and get an AI summary.
For anything you can play or hear right now, the fastest path is a live translator that listens and writes. Verli does this in the browser or as a Mac app: it transcribes the audio as it plays, detects the language automatically, and shows the translation next to the original in real time, for 60+ languages.
Recorded files to translated text
For a file, you have two options. Batch transcription services accept an upload and email you a transcript, which you then run through a translator; that suits long recordings and large volumes.
The simpler route for a single recording: play the file and let a live translator listen. It is exactly the live workflow above, and it needs no upload, which also matters when the recording is sensitive and you would rather not hand the file to another service.
Getting a clean result
Audio quality beats everything else. A clear speaker near the microphone translates well at natural speed; a bad connection, heavy background noise, or a speaker across the room hurts far more than fast talking. If you control the playback, avoid double audio (headphones prevent the transcriber hearing the room and the playback at once), and for files, play them at normal speed rather than accelerated.