Find the best AI-powered app to transcribe your audio

Transcription apps — apps that record and make text transcriptions of the conversation — have been an extremely useful way of keeping records for many of us. This doesn’t only pertain to journalists, who, of course, often keep records of their interviews. When you are caring for a family member who is ill, for example, having a recorded and transcribed record of a conversation with the doctor can be very useful. And when you’re dealing with the insurance company rep afterward — well, ’nuff said.

There are two types of transcription services available online today: one that uses an AI engine and the other that uses human transcribers. The latter is usually much more accurate but is also considerably more expensive. As a result, most individuals use AI-driven services to interpret and transcribe their audio — and admittedly, as AI services have improved, so have the transcriptions. Here is a list of some of the AI-powered transcription services available.

One thing to be aware of: the quality of transcription supplied by these apps can vary widely depending not only on the AI engine the app is using but also on the quality of your audio file. If there are a lot of voices talking at once, if there is a lot of background noise, if the speakers have accents unfamiliar to the AI tool — those can all serve to degrade the accuracy of the transcription. So, a good idea is to try out a transcription service with a typical file to see how well it performs.

And consider which app could be most cost-effective for you. If you only need to upload an occasional file, it could be best to go with either a free version or one of the pay-as-you-go services. If you do regular uploads, then a monthly or annual subscription may work better for you.

Otter was one of the more popular transcription services — well, that is, until August 2022, when it announced downgrades of the services it provides on two of its plans and raised the price on its monthly plan.

That being said, Otter offers a fairly impressive range of services, including the ability to easily record Zoom and Google Meet meetings and organize your transcriptions into folders and your contacts into groups. There is also a separate AI feature that helps with content searches, and each transcription includes an AI-generated summary, including a list of action items and an outline.

As mentioned, there have been a variety of changes to the company’s prices and features. For example, free users no longer have access to all of their past transcriptions — only the last 25 (the others will be archived). You can up to 300 transcription minutes per month with a maximum of 30 minutes per conversation, and you can import up to three audio or video files.

Paying customers who are on Otter’s Pro plan ($16.99 / month or $110.04 a year) once had a monthly allowance of 6,000 minutes of transcribed audio and a maximum of four hours per conversation; these days, they get 1,200 minutes and 90 minutes per conversation; but all their conversations are available, and they can import 10 audio or video conversations per month.

Otter’s Business plan ($30 a month or $240 annually) still has the 6,000 minutes per month / four hours per conversation allowance, along with other features.

Temi is a basic transcription service that is own by the same company that owns Rev — in fact, the first time you go there, it’s likely you’ll find yourself being urged to try Rev first. Once you’ve gotten past that, Temi offers such features as the ability to review and edit your transcriptions, slow down the replay, and export your files into text (Microsoft Word, PDF) or closed caption (SRT, VTT) files. Its mobile apps for Android and iOS allow you to record audio; you can then choose to transcribe it for a straightforward 25 cents per audio minute or upload your own recordings for the same price. New users get the first 45 minutes free.

Rev has been around for a while; until recently, it was mainly available for those who wanted human transcription services. Then the company introduced Rev Max, an AI transcription service that offers 20 hours of automated transcription services and Zoom transcripts for $29.99 per month. (If you pass the 20-hour mark, you’ll be charged 25 cents a minute until your next month begins.) You also get a 5 percent discount on any human-based transcription services. There is a 14-day free trial period, but you have to put in a credit card to get it.

MeetGeek calls itself “an AI meeting assistant.” In other words, its concentration is on transcribing meetings (although it can be used for other audio). It has a free version that lets you create transcripts from audio and video sources — you can record five hours of audio a month and retain three month’s worth of transcripts and one month’s worth of audio. For $19 / month or $180 / year, a Pro version gives you 20 hours of transcription a month, one year of transcript retention and six months of video storage. There are also Business and Enterprise versions. New users get a 14-day trial of the Business plan, which costs $39 a month or $372 a year and gets you 100 hours of transcription a month, unlimited transcript storage, and 12 months of video retention.

Trint’s website makes it obvious that it’s pushing its AI transcription services to creative users; one of the headings on its first page announces, “Our DNA is storytelling.” According to Trint, it can transcribe in over 40 different languages. The Starter 300 plan ($80 / month or $624 / year) lets you transcribe up to 300 minutes per month and do three translations a month, capture audio from its mobile app (iPhone or Android), and edit and share transcripts. The Advanced plan ($100 / month or $720 / year) adds 1200 minutes of transcription along with 20 translations, and the ability to automate workflows. A seven-day free trial lets you sample the Advanced plan.

Sonix offers automated translations in over 49 languages. It includes the usual ability to edit its transcripts, a word-by-word timestamp, and the ability to upload transcripts from other programs and stitch them to new ones. Like many transcription services these days, it has added some AI features such as automated subtitles and summaries. You can export your transcripts in DOCX, TXT, and PDF and export subtitles in SRT and VTT formats. Sonix starts with a pay-as-you-go Standard plan that costs $10 an audio hour (prorated to the nearest minute). There is also a Premium subscription plan ($5 per audio hour plus $22 / month or $198 / year) that adds a number of features and 100GB of storage. New users get 30 free minutes of transcription.

While MeetGeek concentrates on meeting transcriptions, Alice touts itself as a transcription service for journalists. The other services store your transcripts (some with time limits, some without) and let you edit them online, but Alice doesn’t; instead, it sends the audio file and the transcript to your email address and adds it to your Google Drive or Dropbox. It is also easy to use; just tap anywhere on the phone app to start it and swipe to pause. Alice is pay-as-you-go: $9.99 for one or two hours of audio; $99.80 for 20 hours; or $299 for 100 hours. You get the first 60 minutes free and can use it with an iOS app or the web. There is no Android app.

Google Recorder

If you have an Android phone, one of the easiest ways to get a decent transcription is to use Google’s free Recorder app. (If you have a Pixel, you may already have it; otherwise, you can download Recorder from the Play Store to see if it works with your phone.) To start a recording, you simply press a large red button. To pause, you press it again. Smaller buttons on either side let you delete or save the recording. Above the button is the timing of the audio, and above that are two buttons for Audio and Transcript. To see the text, tap Transcript. You can edit the text, search through it (this is Google after all), and share either the audio or the transcript. If you have a Pixel 6 and later, you can enable different labels for different speakers.

Whisper Transcription

OpenAI’s Whisper is an open-source transcription project that is handy to use, especially if you would rather keep your transcriptions off the cloud. There is a Mac app available that makes it easy to install and use if you’re not comfortable with Python and developer tools; if you are, all the better. If you’re using the Mac desktop app, the free version offers several levels of transcription (the slower, the better); a Pro version costs $6.99 a month or $24.99 a year (with a seven-day free trial) and lets you do things such as transcribe podcasts and YouTube URLS. (Whether OpenAI will find itself in a bit of a tangle for training its software using YouTube is another question.)

Update April 11th, 2024, 4:42PM ET: This article was originally published on August 24th, 2022. Since then, several entries have been updated, including info about Otter’s Business plan, while entries for Rev Max, Alice, Google Recorder, and Whisper have been added, and an entry for Scribie has been removed.