I remember a time (not that long ago, actually) when text-to-speech was somewhat of a joke.

However, it's fair to say that things have moved on quite a lot during the last few years, and text-to-speech (especially ones that have public APIs such as Google, Microsoft, NVIDIA, IBM, Amazon, etc.) is now remarkably realistic with almost natural pitch, tone, and so forth.

Considering just how much the technology has evolved, it's no wonder that we're seeing more and more desktop apps that try to leverage it.

One such app is Speech Note, and straight off the bat, it's worth your attention. I promise.

To look at what makes Speech Note so good, let's take a look at the fundamentals. For starters, it's a free and open-source app. It's an AI-powered app - more to the point, it uses the full power of OpenAI  Whisper, alonside multiple other open-source translation and voice libraries and engines (see the full list on the app's GitHub repo).

With that out of the way, here's what it can actually do. It can perform Speech-to-Text (you speak, and the app types), Text-to-Speech (you type, and the app "speaks" back the words), and it can even perform impressive machine translations (including translating text/speech from and to multiple languages).

Other interesting aspects regarding this app include the fact that it can work fully offline (once translation and voice models have been downloaded, of course) and that GPU acceleration is also supported.

Speech Note is, by most accounts, a very impressive application. It becomes even more impressive if you have a decent modern computer with a dedicated GPU.

This is also a good time to point out that this is by no means a lightweight app. Don't get me wrong, it's super efficient when it comes to resource usage, but then again, it will require its fair share of RAM. Oh, and it requires somewhere along the lines of 3 GB of free space.

The only real letdown, and I'm somewhat grasping at straws here, is the GUI. It's not exactly what one might call modern, or stylish. It gets the job done, sure, but it's not much different from that of any other TTS or STT app out there. Definitely not the type of GUI worthy of an AI-powered app (in my honest opinion).