Transcribe interviews on a Mac.

Why local matters for interviews

An interview recording is often the most sensitive file you hold: an off-the-record source, a research participant under an ethics protocol, a subject who agreed to talk to you and no one else. The moment that audio reaches a cloud transcription service you inherit its retention policy, its subprocessors, and its breach surface. Keeping recording and transcription on your own machine removes that whole category of risk: there is no vendor in the loop to trust.

How echo99 records a remote interview

echo99 is built for interviews you conduct over a call (Zoom, Meet, a phone bridge, anything your Mac plays through its speakers):

• Your questions: your microphone, saved as mic.wav and labeled “Me”.
• Their answers: the other side’s audio, saved separately as system.wav; on-device diarization labels each voice “Them-1”, “Them-2”, …, and speaker identification can put a name to a voice you interview again.

Transcription runs on-device afterwards into transcript.txt and transcript.json, with speaker labels and timestamps. The transcript is searchable, ready to review for quotes, and never uploaded.

Working with the transcript

• Play any transcript line on its own, or click it to move full-call playback to that moment, so you can check a quote against what was actually said.
• Use Export as Markdown… in the Recordings window to take the transcript into your notes or draft.
• The default speech model handles 25 European languages, Czech included. The technology page lists them.

In-person interviews

echo99 separates the two sides of a call because the far side comes through your Mac’s audio output. For an in-person interview, everyone in the room is picked up by your microphone, so if echo99 records it, both voices land on the “Me” track and are not told apart.

If you want the voices separated, record the interview on your phone (for example with Voice Memos) and import the file into echo99 afterwards. An imported recording is treated as one mixed track, and diarization labels its voices “Them-1”, “Them-2”, and so on, which you can then name.

Consent comes first

echo99 removes the technical and cloud barriers, not the ethical and legal ones. Recording-consent laws vary by jurisdiction, and research and journalism carry their own standards. echo99 doesn’t announce itself, so getting and documenting consent is on you. Ask at the start, and keep your recording practice within your obligations.

New to echo99? See the getting started guide.

Related guides: Record a podcast, Local transcription