What is noScribe?
noScribe is a desktop app for turning audio recordings into text, built mainly for researchers and journalists who work with interviews. It combines the Whisper speech recognition model with pyannote speaker detection, so the transcript shows who said what.
Everything runs on your own Windows, Mac or Linux computer, which means recordings never leave your machine. That matters for confidential interviews and research data. noScribe supports around 60 languages and includes its own editor for checking and correcting the transcript against the audio. There is also a Docker option for running it on a server without a window. Note that the noscribe.ai domain is not connected to this project.
Main Features
- Local transcription with Whisper, no upload needed
- Automatic speaker identification
- Built-in editor to review and correct transcripts
- Around 60 languages supported
- Windows, macOS and Linux versions
- Docker option for batch jobs on servers
How to Get Started with noScribe
- Go to noscribe.de and download the version for your system.
- Install the app and open your audio file.
- Choose the language and whether to detect speakers.
- Start the transcription, then check the result in the built-in editor.
noScribe Pricing
noScribe is completely free and open source under the GPL-3.0 licence. The project says it will always be free and accepts optional donations.
Who Should Use It?
- Qualitative researchers
- Journalists transcribing interviews
- Students working on theses
- Anyone handling confidential recordings
- Podcasters who want offline transcripts
noScribe Alternatives
- TurboScribe: Online AI transcription with a generous free allowance.
- Happy Scribe: Transcription and subtitle platform with AI and human options.
- Superwhisper: AI dictation and transcription app for Mac.
Frequently Asked Questions
Is noScribe free?
Yes. noScribe is free and open source, with no paid plan. Donations are welcome but optional.
Does noScribe upload my audio?
No. Transcription runs entirely on your own computer, so your recordings stay on your device.
Can noScribe tell speakers apart?
Yes. It uses pyannote to detect different speakers and labels them in the transcript.