Supported audio formats
You can upload MP3, WAV, M4A (the format used by iPhone Voice Memos), AAC, FLAC, OGG/Opus and audio-only WebM files. Before transcription, the audio is converted to 16 kHz mono, which is what speech recognition models work with, so a high-bitrate file won't improve accuracy but will take longer to upload.
Getting good results from recordings
Interviews and phone calls
Place the recorder between speakers, or record each side separately if your app allows it. Only record calls when everyone has agreed — consent rules vary by state.
Lectures and presentations
Sit near the speaker or record from the room's sound system. Distant, echoing audio is the most common cause of missed words.
Podcasts
Upload the final mixed episode. Intro music is fine, but music playing under speech can reduce accuracy in those sections.
What you get
The transcript appears line by line, each with its start time, so you can find the moment in the recording quickly. Download it as plain text with or without timestamps. If you want captions for a video or audiogram, the Subtitle Generator creates SRT and VTT files from the same kind of recording.
Limits to know about
- Each recording can be up to 200 MB and 60 minutes long.
- The transcript is written in the language that is spoken. It doesn't label speakers and isn't translated.
- If your recording is a video file, upload it to Video to Text instead. You don't need to extract the audio first.