I gave this blog a voice

This week I added a "listen to this post" button to every post on this blog, including this one.

It uses ElevenLabs, and it reads in a clone of my own voice, generated once when a post is written and served as a plain audio file after that. No server calls the API while you are reading. It just plays back what was already made.

Why bother? Because not everybody wants to read. Some of you would rather have this in the background while you are doing something else. That is a fair way to take in three minutes of writing, and now you can.

Is it good?

Mostly. It is pretty good.

The pacing is close to how I actually talk. It handles the plain language I write in without stumbling. Sped up a notch, it sounds less like a machine reading my words and more like me in a hurry, which is closer to the truth anyway.

It is not perfect though. Odd words trip it. A stray line of code or a acronym read out loud sometimes sounds strange, so I am stripping that kind of thing out of what gets sent to the voice rather than trying to fix it after the fact. It will keep needing small adjustments like that as I find them.

What I am actually doing here

I am not just turning this on and walking away. I am trying to learn about the process and its quality and its limitations, the same way I would with any new tool before I hand it to somebody else.

That means listening to my own posts back, more than once, and noticing where it is off. It means understanding what this costs in practice, not just in theory. It means deciding, post by post, whether the audio version is actually worth having or just a novelty that wears off.

So far the answer leans toward worth having. Not perfect. Pretty good. Improving.

Be Better - Use your voice.