No, okay, I get it: there are situations where it's easier to quickly record an audio. When you're driving, for example. But how often does it happen that instead of expressing a thought in one or two sentences, your interlocutor sends you a voice message lasting a couple of minutes, where the lion's share is "um" and "well, in short, like".
Good thing that the same Telegram allows you not to listen to all that, but simply press a button and get a transcript. But the case of "dictate into text and send" remained unresolved for me for a long time. Because typing long texts is always friction: sit down, gather yourself, start. And dictation removes the initial barrier.
Yes, voice input exists, but it usually works terribly. Even if you try to dictate with punctuation marks. The result is usually something like "that's the way it is period and comma however...".
But recently I stumbled upon Aqua Voice, which has — surprise! — a neural network under the hood. And thanks to that, you get a polished transcript.
First, the periods and commas are in order. Second, if you misspeak, you can immediately correct yourself, and in the text you will only see the corrected version. Third, you can use a prompt to set processing settings for your audio, for example, ask to break the text into paragraphs. Or automatically remove filler words like "in short" at the beginning of a sentence.
I was skeptical, but after a few weeks of use, I admit — this is a must-have. Now you don't have to go on and on, typing long and tediously. You can just dictate and get a decent result.
This thing really shines during vibe coding. Dictated what you need — got the result. It's like working with a live person, in short.
In general, if you, like me, hate typing long texts — it's worth a try. Upon registration, you get a month of free use. After that — $10 per month.
Oh, yes, there's also a referral link. If someone registers through it, you get another free month. Here's mine — click.
Comments
0No comments yet.