Transcribing Audio: Handy Tools and Privacy

Transcribing audio can be done in minutes and even for free, as long as you use the right tools. This article outlines four simple steps to convert an audio recording into text. The process includes uploading the file, selecting the correct language, starting the transcription, and checking the outcome. It’s important to thoroughly review the result, as automatically generated transcriptions often contain errors.

There are various free tools available for audio transcription. Microsoft Word, ElevenLabs Scribe, Whisper, and Amberscript all offer solutions, but with different limitations. For instance, Word provides 300 minutes per month if you have a Microsoft 365 subscription, while Scribe can process up to 10,000 credits per month. Whisper is an open-source option with no limit, but requires some technical knowledge for installation. The article also specifically covers how to effectively use ElevenLabs Scribe to convert a recording with multiple speakers into text.

When transcribing using ElevenLabs Scribe, the process begins with creating a free account, after which you can upload your file. The tool offers detailed features such as timestamps and recognition of various speakers. After transcription, it's essential to review the text for discrepancies, especially concerning names, technical terms, and the order of speakers. The article further provides advice on privacy issues if you are processing sensitive audio.

The use of ChatGPT for transcription is also discussed. While there are capabilities for converting audio to text, the results are less reliable than those from specialized transcription tools. Moreover, the article addresses the different types of transcription, including verbatim and summary transcription, and gives insights into the security of audio with online tools. It is crucial to know that sensitive information is best transcribed locally using software like Whisper.

Read the full article from AI Insider.