Murmur for Researchers

Manually transcribing a one-hour interview takes four to six hours. Run ten, twenty, or more interviews for a study and you spend weeks just typing. Automatic transcription speeds this up enormously — but most tools upload your audio files to external servers.

For research interviews, that's a particular problem. Participants share personal experiences, sensitive topics, sometimes under assurances of confidentiality. Ethics boards and data protection officers have legitimate questions when these recordings are processed through cloud services.

Record interviews directly in Murmur — via the built-in microphone, an external recorder, or system audio for remote interviews over Zoom or Teams. Or drag and drop existing recordings into the app.

Speech recognition (Parakeet ASR) runs locally on your Mac and transcribes in 25+ languages with automatic detection. For interviews with multiple participants, speaker diarization automatically assigns statements to different speakers.

After transcription, a local language model generates a summary — key themes, recurring patterns, central statements. Click a timestamp to hear the corresponding moment in the audio.

When you assure participants of confidentiality, that assurance should hold up. With Murmur, recordings and transcripts never leave your computer. There's no third party in the processing chain, no servers where audio files are temporarily stored.

This significantly simplifies the argument to ethics boards and data protection officers. Instead of analyzing the privacy policies of a cloud provider, you point to local processing and local encryption.