noScribe

Name: noScribe
Author: noScribe

Free plan LinuxMacWindows

What noScribe is

noScribe is a free, open-source desktop app for transcribing interviews and other sensitive audio with local AI models. Audio and transcripts remain on the user's computer rather than being sent outside. It uses Whisper-based transcription and pyannote for speaker recognition, supporting around 60 languages. The editor aligns transcript text with audio playback for review and correction; users can mark pauses and overlapping speech, add timestamps, and choose whether to include filler words. The app handles almost any audio or video format and exports formatted HTML, plain text, or WebVTT subtitle files. Transcripts can be opened in MS Word or LibreOffice and imported into QDA software such as MAXQDA or QualCoder. The maker estimates that a one-hour interview takes about one to three hours to transcribe, with NVIDIA CUDA acceleration able to reduce the time. noScribe is available for Windows, macOS, and Linux, though Linux binaries are not yet available for version 0.8. It is licensed under GPL-3.0; the maker says transcripts may be used commercially without restrictions.

Who it suits

It suits researchers, journalists, and others working with interviews or sensitive recordings who want local transcription. The project also describes use in medicine, companies, government agencies, and NGOs.

What it does well

Audio and transcripts stay on the user's computer.
Supports around 60 languages and speaker recognition.
Editor aligns text with audio playback.
Exports HTML, plain text, and WebVTT.
Maker permits commercial use of transcripts.

What to know first

One-hour interviews may take one to three hours.
Linux binaries are unavailable for version 0.8.
macOS 14 or later is required.
Windows 7 and 8 are unsupported by current versions.

Plans

Plan	Billed and limits	Price
noScribe	Free and open source · noScribe is free and will always remain so	Free

As noscribe.de lists them, read 11 Oct 2026. The maker's own page has the final word.

Getting started

Visit the noScribe website and choose a download route for your operating system.
Use Windows 10 or macOS 14 or later; for Linux, choose version 0.7 or manual installation if using the version 0.8 download page.
Open an audio or video file for transcription.
Review and correct the aligned transcript in the editor.
Export the transcript as HTML, plain text, or WebVTT.

Verdict

noScribe offers local transcription, speaker recognition, and an editor suited to reviewing interview recordings. Check the operating-system requirements and allow for the maker's estimated processing time.

Questions

How much does noScribe cost?

The listed noScribe plan costs nothing. The software is free and open source under GPL-3.0.

Which operating systems does it support?

noScribe lists Linux, macOS, and Windows. The FAQ says it runs on Windows 10 and requires macOS 14 or later; current versions no longer support Windows 7/8 or older Intel Macs.

Does noScribe identify speakers?

Yes. It uses pyannote for speaker recognition and Whisper-based models for transcription.

Does it send recordings to an online service?

No. Transcription runs locally on the user's computer, without sending audio or transcripts outside it.

Can transcripts be used commercially?

The maker says transcripts created with noScribe may be used commercially without restrictions. Modifying the software is subject to GPL-3.0 terms.

How long does transcription take?

The maker estimates about one to three hours for a one-hour interview. NVIDIA CUDA acceleration can reduce processing time.

Facts from noScribe's own pages

Command line: Since version 0.7, noScribe provides a command-line interface that can be integrated into scripts. · noscribe.de, 11 Oct 2026
Commercial use: The maker says users may use transcripts commercially without restrictions, while modifying the software is subject to GPL-3.0 terms. · noscribe.de, 11 Oct 2026
Editing: A dedicated editor lets users review transcripts while aligning the text with audio playback. · noscribe.de, 11 Oct 2026
File support: The app can process almost any audio or video format. · noscribe.de, 11 Oct 2026
Free plan: Yes · noscribe.de, 20 Sept 2026
Headquarters: Frankfurt am Main, Germany · noscribe.de, 28 Sept 2026
Input and output: It supports almost any audio or video format and can export HTML, plain text, or WebVTT files. · noscribe.de, 11 Oct 2026
Integrations: Transcripts can be opened in MS Word or LibreOffice and imported into QDA software such as MAXQDA or QualCoder; WebVTT can be used with EXMARaLDA. · noscribe.de, 11 Oct 2026
Intended users: The project began for qualitative social research and is also used in journalism, medicine, companies, government agencies, and NGOs. · noscribe.de, 11 Oct 2026
License and use: noScribe is released under GPL-3.0, and the maker says there are no restrictions on commercial use of transcripts created with it. · noscribe.de, 11 Oct 2026
Linux availability: The download page says Linux binaries are not yet available for version 0.8 and directs users to version 0.7 or manual installation. · noscribe.de, 11 Oct 2026
Models: The software uses Whisper-based transcription and pyannote for speaker recognition, and supports around 60 languages. · noscribe.de, 11 Oct 2026
Open source: noScribe is free and open source under GPL-3.0. · noscribe.de, 11 Oct 2026
Output formats: noScribe can create formatted HTML, plain text, or WebVTT subtitle files. · noscribe.de, 11 Oct 2026
Performance: The maker estimates that transcribing a one-hour interview takes about one to three hours, with NVIDIA CUDA acceleration able to reduce processing time. · noscribe.de, 11 Oct 2026
Privacy: Transcription runs locally on the user's computer without sending audio or transcripts to the outside. · noscribe.de, 11 Oct 2026
Product: noScribe is a free, open-source desktop app for transcribing interviews and other sensitive audio data with local AI models. · noscribe.de, 11 Oct 2026
Purpose: noScribe is a desktop app for transcribing interviews and other sensitive audio data with local AI models. · noscribe.de, 11 Oct 2026
Research features: Users can mark pauses and overlapping speech, insert timestamps, and choose whether filler words are transcribed. · noscribe.de, 11 Oct 2026
Speaker identification: Yes · noscribe.de, 20 Sept 2026
Support: The maker recommends GitHub Issues for help and provides email contact for urgent cases. · noscribe.de, 11 Oct 2026
System requirements: The FAQ says noScribe runs on Windows 10 and requires macOS 14 or later, while Windows 7/8 and older Intel Macs are no longer supported by current versions. · noscribe.de, 11 Oct 2026
Transcript editing: An included editor lets users align audio and text to review and correct transcripts. · noscribe.de, 11 Oct 2026
Transcription: The software uses Whisper-based models and supports speaker recognition and around 60 languages. · noscribe.de, 11 Oct 2026