Transcribe audio and video files you already have.

How to import a recording

Import becomes available once you have finished setting up echo99 and the speech and speaker models selected in Config → Models are installed. Until then the Import button and the Import Recording… menu item are greyed out.

1. Open the echo99 menu in the menu bar and choose Import Recording….
2. Drop an audio or video file into the window, or click Choose File….
3. Review the recording name and date. If you use the offline speaker model, you can also set Other speakers to Auto or a number from 1 to 4. Every voice in an imported file counts as another speaker, so a two-person interview is 2.
4. Click Import. echo99 converts the audio locally, adds it to Recordings…, and queues it for transcription.

Supported audio and video formats

echo99 imports formats that macOS can decode. Verified common formats include WAV, MP3, M4A, FLAC, MP4, and MOV. For video, echo99 uses the default audio track; a video without audio cannot be imported. Files that macOS cannot decode are rejected cleanly, without changing the original file.

What happens to the file

• The source file stays where it is and remains unchanged.
• echo99 creates a local recording using the name and date you chose, then applies your current transcription, vocabulary, silence, and speaker settings.
• Audio conversion, transcription, and speaker labeling run on your Mac; no account or cloud upload is required.
• Playback, speaker naming, re-transcription, and Markdown export work just like they do for a call recorded by echo99.

Why an imported recording has no “Me” track

A live call lets echo99 preserve your microphone and the other side separately. An existing file already combines its voices, so echo99 treats it as one mixed track. Speaker separation initially labels its voices “Them-1”, “Them-2”, and so on; there is no separate microphone track labeled “Me”. You can name speakers from the transcript afterward, and echo99 can recognize those familiar voices in later recordings.

Continue in your recordings library

When conversion finishes, choose Show in Recordings. The imported item is ordered by the recording date you selected and stays available for transcript playback, speaker edits, re-transcription with another model, and export.

Limits

• A long file takes a while to transcribe, and an Intel Mac is slower than Apple silicon.
• The default speech model handles 25 European languages; the technology page lists them.

Related guides: Local transcription, Speaker identification