The App That Actually Understood Both My Languages

The idea of talking instead of typing had been sitting in the back of my mind for a while. Walking, thinking, and capturing ideas at the same time — that sounded like a real productivity unlock. The problem was finding a speech-to-text app that actually worked.

For a bilingual Thai-English speaker, the bar is embarrassingly low — and most apps still fail to clear it.

They pick one language and ignore the other. Or they output something random entirely. I switch between English and Thai constantly — sometimes mid-sentence, sometimes talking to myself in both at once. That is just how I think. Most apps could not keep up.

Found It Early

Voicenotes launched in April 2024, built in 49 days by a small husband-and-wife team. I found it shortly after, while researching speech-to-text tools. I was not looking for a notes app specifically — I was looking for something that could keep up with how I actually talk.

I got in early. Early enough that the product was still raw and finding its footing. That said, even then, it did the one thing most apps could not: it handled English and Thai in the same recording, in the same sentence, and got it right.

That was enough to keep me around.

The Language Problem Nobody Talks About

If you speak one language, you probably have not noticed how badly speech-to-text apps handle code-switching. But if you regularly mix two languages in the same thought, you know exactly how broken the experience gets.

Pick one language: it ignores half of what you said. Try to force both: it guesses wrong, sometimes outputting a third language entirely.

Voicenotes does not do that. It transcribes what you actually said, in the language you said it. English sentence, Thai sentence, English phrase mid-Thai thought — it follows along. That accuracy is the foundation everything else sits on. Without it, none of the other features matter.

How It Actually Fits In

My original use case was simple: speak instead of write. I started using it like a journal — instead of typing out my thoughts, I talked into it. Structure my thinking, process ideas, dump what was in my head. Same purpose, different input.

From there the use cases expanded on their own:

• Driving — I cannot type behind the wheel, but I can talk. Ideas that used to evaporate on the road now get captured.
• Walking and thinking — There is something about moving that loosens up my thinking. Walking while recording pulls out more ideas than sitting at a desk does.

To be clear: this is not a replacement for writing or typing. It is not trying to be. It is another tool in the toolkit — one you reach for when talking is faster, or when writing is not an option, or when you just want to think out loud and have something waiting for you on the other side.

The Meeting Recording Feature I Almost Never Tried

When Voicenotes added meeting recording, I barely noticed. I was already paying for another tool that records online meetings, and it worked fine. I did not need a second one.

What I did not pay close attention to was that it also records offline meetings — in-person conversations, not just calls. That was different.

A good friend reached out asking for a recommendation: an app that records offline meetings and does it well. I pointed him to Voicenotes. Told him I use it and it is solid — even though I had never actually tried the meeting recording feature myself.

A few days later, he came back:

"dude
this voicenotes - amazing shit
thanks alot"

That was all I needed. I went and tried it myself.

It delivered. The transcription was accurate and clean. And for online meetings, it does not need to join the call — no bot appearing in the participant list, no notification sent to the room. It just runs quietly in the background and records what you hear. That level of discretion is genuinely useful. And for me, a smaller but real benefit: it removes the low-level anxiety of realizing partway through a meeting that nothing is being captured.

There Is Also an AI Layer

Two features worth knowing about, even if you do not use them constantly:

• Search across all your recordings — find something you said weeks ago without scrolling through everything manually
• Prompt to summarize, clean up, or extract from a transcript — you give it an input, it gives you a usable output

I do not use these as a core part of my workflow. The recording and transcription value is already enough on its own. But the AI layer is there, it works, and it adds up when you do reach for it.

Worth It

Voicenotes has earned a permanent spot in my toolkit. Not because it does everything — it does not try to. But what it does, it does well. And for anyone who lives between two languages and has given up on speech-to-text because it keeps getting it wrong: the accuracy alone is worth trying it.

Sometimes one thing done right is enough.