Tools for Automatic Speech Recognition

noScribe is a free open-source tool for automatic transcription and subtitling. It works offline with Whisper (OpenAI) and offers: Automatic timecodes and speaker labels (S00, S01, etc.) Marking of pauses and sounds Built-in text editor to correct transcriptions Click here for an overview of the installation page and developments.

MacWhisper converts audio files to text using OpenAI’s Whisper technology. Everything happens locally, keeping your data safe.

• Privacy guaranteed – All transcriptions are processed locally.
• Versatile file formats – Supports mp3, wav, m4a, and mp4.
• Handy export options – Export subtitles (.srt, .vtt) or text (.csv).
• Over 100 languages – Automatic recognition and transcription.
• Smart extras – Removes filler words such as “uh” and “um.”
• User-friendly – Search and highlight text, synchronize audio.

For €59, you get access to more powerful Whisper models for more accurate transcriptions.

DoblAI is an online tool for transcribing, translating, and dubbing multimedia. It is free, platform-independent, and offers automatic speaker recognition for clear transcripts.

• Get started right away – No installation required, works in your browser.
• Speaker recognition – Automatically recognizes different voices.
• Expandable options – Paid features for faster processing and extra control.
• Privacy – Files are encrypted and comply with GDPR.
• File format – Exports .srt with speaker colors.

Ideal for accessible, flexible transcriptions.

Subtitle Edit supports speech recognition via two systems:

• Vosk/Kaldi: an older method
• Whisper: modern AI technology with higher accuracy
• Whisper engines: Some (such as OpenAI, WhisperX, CTranslate2) require Python. Others (such as Purfiew’s Faster Whisper-XXL and CPP) work without Python.
• Download Whisper via the menu: Video → Audio to text (Whisper)
• Model selection: Select large-v2 under Choose model for the best results
• Settings: Advanced options and post-processing can be found under Audio to text
• Online use: Subtitle Edit can also be used entirely online

aTrain is a powerful tool for automatic transcription, including speaker recognition. Everything happens locally on your device, fully GDPR-compliant.

• Privacy: No data storage in the cloud
• Speaker detection: Recognizes who is saying what
• Fast: Support for NVIDIA GPU
• Compatible: Works with MAXQDA, among others
• Windows: Via Microsoft Store or BANDAS Center (from Windows 10 v17763)
• macOS: Beta version 1.2.4 via alternative installation

Riverside effortlessly converts audio and video into text, directly in your browser. No downloads, no installation, free to use, and supports over 100 languages.

• Unlimited uploads – Process MP3, WAV, MP4, and MOV files without restrictions.
• Convenient export options – Export subtitles (.srt) or text files (.txt).
• Super simple – Start the transcription in just a few clicks.
• Privacy – Files are uploaded online.
• Processing time – Depends on file size and server traffic.

Ideal for quick, easy transcriptions without software.

The CLARIN Transcription Portal is a user-friendly tool for automatic speech recognition developed by LMU Munich. It is designed for academic use and supports multiple ASR services.

• Multilingual – Supports transcription in multiple languages.
• Editable – Correct and edit transcriptions in the OH Portal.
• File format – Only .wav files are supported.
• Accessibility – Free for academics with a CLARIN account.

The portal switches between audio files and ASR services, with export and editing options for transcriptions.

The CLARIAH ASR tool, developed by Radboud University, offers a powerful online service for automatic transcription and speaker recognition. Using WhisperX (speech recognition) and PyAnnote (speaker diarization), the tool delivers accurate and clearly structured transcriptions in multiple languages.

• Multilingual support – Suitable for transcription in different languages.
• Speaker recognition – Automatically recognizes and separates different speakers within a single recording.
• Secure processing – All data remains the property of the user and is only accessible to technical staff.
• Data storage and management – Projects can be deleted by the user and are automatically deleted after 30 days.
• No encryption, but protected – Data is not shared with third parties and is only used for transcription purposes.

This service offers a flexible and reliable solution for researchers who need transcripts with speaker information. Please note: save transcription results locally in a timely manner, as data is deleted after 30 days.

ELAN is a free tool from The Language Archive for manual, word-by-word transcription and linguistic annotation. It supports complex, multi-layered analysis.

• Flexible use – Create annotations such as transcriptions, memos, codes, or key moments.
• Word-by-word – Detailed segmentation possible.
• Multi-layer transcription – Use tiers for different speakers or analysis categories.
• Not automatic – Transcription is manual; ASR linking is possible via external tools.

Ideal for linguistic research and those who want to annotate in depth with structure and precision.

oTranscribe is a user-friendly, free tool for manual transcription via a simple web interface.

• Web version – Use the tool directly in your browser.
• Audio and video files – Supports uploading your own files or YouTube videos.
• User-friendly – Combines a text editor with a simple media player and keyboard shortcuts.
• No timecodes – Are not automatically added to the transcription.
• Privacy – Data is stored locally in your browser.

Available as a web tool and Chrome app.

• Cross-platform – Available for Windows, macOS, and Linux
• Free and open-source – Completely free to use
• Manual transcription with timecodes – Synchronize subtitles accurately to audio
• No media player – Not intended for playing video and audio simultaneously
• Export as .srt – Use Export to save subtitles