Record a text note with your voice and AI
Datum: 2023-06-08 12:09
Siri doesn’t understand me!
How many times have you tried dictating a quick voice note into your phone to capture the essence of a meeting only to then get back a text transcription of what you said that doesn’t match up with what you said?
I think I’m fairly articulate, but at least I have had to spend unnecessary time cleaning up the note I made afterwards. Or, I have later pondered long and hard on what I meant by “barbecuing” and “semi-trailer”. I can’t remember myself and my customer talking about that.
For you who prefer listening to reading, this post is also available as an episode of the “Done!” podcast:
Quickly caught on the go!
The idea of being able to quickly record a message and get it in text form is brilliant. If it works, you can capture an idea in a flash. You quickly document what you agreed on. Without worries, you add what you promised to your to-do list — even if you’re on the go.
A solution where you are understood
But, at least I haven’t been able to make it work — until now. A few weeks ago though, I did find a solution that has worked better than anything else I have tried in this direction. I can’t help but tell you about it. (Since development in this field is progressing rapidly now, there surely are other solutions — and any day now OpenAI’s mobile app will be to everyone, too — so consider this solution as one among many that works.)
I record a voice note in the phone and a short while later I get an email where what I said is written down — with astonishing precision. Especially fascinating is that the text reflects the meaning of what I said — not exactly every “hm”, mistake and repetition I made in the voice recording.
The components involved in the solution are:
- The RecUp app (iOS)
- The storage service Dropbox
- The automation service Zapier
- The speech recognition module Whisper
- The email platform Gmail.
Here is how it works
When I record a voice note in RecUp (1), it gets saved as an mp3 file in Dropbox (2). When Zapier sees a new mp3 file in the voice notes folder, it fetches the file (3) and sends it through the OpenAI plugin to Whisper (4), which responds back with the text it has understood, and then Zapier writes an email in Gmail which it sends to me (5).
Do this:
- First, get RecUp (iOS) or another voice-recording app where you can save your recordings to Dropbox (there are plenty of Android apps to choose from).
- In RecUp, set up which Dropbox folder you want the file to be saved in.
- In Zapier, create a ‘zap’ that:
- picks the mp3 file from the Dropbox folder when it’s saved there,
- creates a transcription using the OpenAI plugin (you’ll need to have a subscription to OpenAI’s API first)
- sends you an email from your own Gmail account (or other email platform such as Outlook) with the transcribed text in it.
- Test it out — and be amazed at how accurately Whisper understood you. It’s something else than Siri, that’s for sure!
More Time for the Essential
If you use this solution or a similar one to quickly make a voice recording, you need to spend less time taking notes after meetings and conversations. You capture valuable ideas on the go that would otherwise have disappeared into eternity. The many small streams of minutes you save become a large river of hours that you can dedicate to what is most important for you and your work.
What is your solution?
Have you found another, similar solution to the need for quick voice recordings to text? Tell me about it! I collect solutions like this, so I am eager to hear what you’ve found.
(Have you heard about the other things that I’ve been getting help from AI with so far?)
Book a talk about structure for your next kickoff
Simplifying tips and tricks like this are exactly what I share in my talks. If you'd like me to come to your next kickoff event, get in touch now!