AudioPen is a wonderful tool. If you haven’t encountered it yet, it’s a simple product that claims to “go from fuzzy thought to clear text fast.” And it does just that. You record yourself talking, and it translates your speech into text in a style you can define. Think of it as classic dictation software with a twist—generative AI can be applied seamlessly at the end of the recording.
There’s a free version with a time limit, and a paid tier offering premium features like creating your own styles. I’ve made several styles myself and would be happy to share them. AudioPen has been a huge productivity booster for me, especially for this blog. In fact, I used it extensively for this very post!
My current writing process is something of a three-way dance between my notes, AudioPen, and ChatGPT. I often create an initial outline in my notes, dictate it into AudioPen (preserving the spirit of the original), and then have ChatGPT do some light editing. Admittedly, it doesn’t always work perfectly—I don’t want my writing to sound AI-generated, which it sometimes does. But for now, I’m content with putting my thoughts out there and telling myself I’ll edit these posts when I have more time.
AudioPen isn’t limited to blogging, though. It’s great for thank you notes, lengthy text messages, Slacks, or emails when you know what you want to say but dread turning it into something coherent. With AudioPen, a task that might have taken 10 minutes becomes a two-minute breeze. Need to write a speech? AudioPen has you covered.
It can even help with note-taking—say you’ve watched a video but didn’t take notes while viewing. You can use AudioPen to quickly record what you learned and maybe even store it in your Obsidian vault.
Now, I suspect that AudioPen’s core functionality will soon be commoditized, if it hasn’t already. Speech-to-text with generative AI is likely to become a standard feature in operating systems.
And like other AI tools, third-party options often involve communicating over a network and storing data on their servers. So, it’s essential to be cautious and discerning about what information you put out there. As far as I understand, AudioPen stores all the sound files and text.
Plus, you might be aware that you can download and run local speech-to-text models, like Whisper, on your own hardware. So there are certainly alternatives out there.
That being said, I highly recommend giving it a try in a world where new AI products emerge every day. It’s one of the few that has become part of my daily workflow. For me, it’s right up there with GitHub Copilot and ChatGPT as the essentials. Plus, it’s created by an indie hacker, and I love supporting those types of creators. In my opinion, it’s well worth the investment.