noScribe is a free, open-source desktop app with 2,200+ stars on GitHub that transcribes interview audio automatically with speaker identification — built specifically for qualitative researchers and journalists, and running entirely on your own computer.
Transcribing interviews is slow, expensive and often involves uploading confidential recordings to cloud services. noScribe, actively maintained through 2026, removes all three problems by standing on the shoulders of giants: it wraps OpenAI’s Whisper (via the faster-whisper implementation) for speech recognition and pyannote for speaker diarization in a friendly graphical interface, then adds a built-in editor where you review, verify and correct the transcript with the audio at hand. It distinguishes between different speakers, understands around 60 languages, and runs completely locally on Windows, macOS and Linux — no cloud, no account, and no risk to interview confidentiality, which is precisely what research ethics boards and newsrooms need. The developer is emphatic that the real noScribe is free and always will be, and warns users away from unrelated commercial sites trading on its name.
noScribe is completely free and open source under the GPL-3.0 license — the developer states it will always be free. Download it from the official project channels and run it locally; there is no paid version.
Qualitative social researchers, journalists, oral historians and students who transcribe interviews regularly and need accurate, speaker-labeled transcripts without cloud services or per-minute fees.