As a journalist and blogger, I do a lot of interviews: at press events, over the phone, and now via Zoom. While I once relied solely on a notebook, such as when I spent three days on the set of Trading Spaces, I now record everything.
Transcribing is my least favorite part of the journalistic process. I loath it—and not just because I must listen to my Muppet voice. It takes so much time, especially because I want the transcript to be accurate and complete so that any quotations I draw from it are also accurate.
Interviewing Survivor’s cast on location, before the game began, left me with interviews I was so excited about—and which I let sit on my hard drive because the work of manually transcribing them felt overwhelming.
When I heard about Rev, I started using that for longer interviews, especially when I needed quick turnaround. I was glad to be paying talented people to do that work for me. Alas, I found Rev wasn’t always accurate, so I was spending a lot of time cleaning up transcripts, and not just for expected things such as proper names.
Is there a better transcription than Otter?
When I first heard about Otter, which uses AI to transcribe in real time, I found it to be far more accurate than Rev. As a bonus, its transcript interface made it easy to identify speakers and then go back and forth. I was in love, and so were many other journalists, academics, bloggers, podcasters, and those who need to transcribe.
I paid $99 a year for Otter’s pro plan, and it was totally worth it. Then, last fall, Otter changed their plans. I don’t mind companies adjusting prices. But what Otter did was drop features from its pro plan: the number of hours that could be transcribed went from 100 to 20, and they reduced uploads of audio files to just 10 when that was previously unlimited.
As New Yorker writer Ben Taub tweeted, that plan change “renders it useless for journalists.”
If those of us on the Pro plan wanted more uploads and minutes, we’d have to upgrade to the next plan, for $240 a month. I was so annoyed I tweeted:
“Hi, we’re @otter_ai, and we do one thing extremely well that is worth paying for—so here are a bunch of new features you don’t use, and we’ll take away the ones you do, so you can pay even more if you want those back!!”
Maybe Otter’s internal data showed that people like me were a tiny group and their real clients were corporations looking to transcribe meetings, hence all the new features they added that were worthless to me, like meeting notes.
So I began searching for something else to transcribe interviews.
There are a bunch of options, but most were far more expensive than Otter—which may explain why they didn’t mind raising their prices. I considered going back to Rev, which incorporated Temi, so now it offers automated transcripts for 25 cents per minute.
I found Alice, which is really impressive, as they’ve designed an easy-to-use app for journalists reporting in the field.
But I settled on MacWhisper. Here’s why.
MacWhisper: the Otter.ai alternative I chose for transcripts
MacWhisper uses OpenAI’s Whisper engine, and I’ve found that to be incredibly accurate—even recognizing reality TV show titles and cast member names.
For those worried about companies using our creative work to make their products better, I was especially attracted to what MacWhisper says in its description: “All transcription is done on your device, no data leaves your machine.“
The app’s features are amazing: you can upload an audio file; record live, as Otter does; or record another app’s audio, such as FaceTime or Zoom. And it allows for easy transcript exports in multiple formats.
Oh, and it’s free!
MacWhisper does have a pro version, which I quickly upgraded to, as it was only €20 to add features such as access to other language models, and podcast transcribing that combines audio for each participant.
That’s a one-time upgrade; MacWhisper says that includes “access to all current and future features.” (MacWhisper is also in the Mac App Store, though if you buy it there, the Pro version is $27.99/year.)
For me, it was easily worth €20—about US$22—even if I’d never use any of the pro features.
The only downside is that its transcript interface isn’t quite as shiny as Otter’s, which has color-coded speaker names. That was imperfect, though, as Otter often didn’t notice when someone interjected or when speakers changed rapidly.
I’ve found that MacWhisper gives me exactly what I need to pull quotes for stories.
If you have a Mac and a need to transcribe audio live or from recordings, give it a shot. I hope it works as well for you as it has for me.