Why I made murmrrr
I'm Elliot. I'm 14, I live in England, and I built this because the dictation apps I wanted to use cost more than I had.
It started with a course, then games
I did a three-day AI course meant for adults. It covered what AI actually is, how Claude works, how to code with it, and how to deploy what you make so other people can use it. That last part stuck: it's one thing to make something, and another to put it out in the world.
After that I got into coding as a hobby by making little games with Claude, the AI from Anthropic. You describe what you want, it writes the code, you test it, you argue with it about why the jump doesn't work, and eventually you've got a game your friends will actually play. People call it vibe coding. I just called it fun.
Then I wanted to talk instead of type
I type slowly and I talk fast, like most people. I tried the big dictation apps and they were good, but they cost about £140 a year, forever, and most of them send everything you say to a server somewhere. On pocket money, a subscription like that isn't a thing you have, it's a thing you cancel.
So I thought: I've built games. How hard can a dictation app be?
Harder than a game, it turns out
Getting a computer to hear words is only the start. The work was everything around it: making it run on your own computer so nothing is uploaded, getting it to type into any app, including games, making it put in full stops and commas and lay out a list when you say a list, keeping it out of the way in a little bar, and making it survive Windows and Mac updates. Claude wrote a lot of the code. I decided what it should do, tested every bit of it, and fixed the bugs, a lot of them late at night.
What I'd like you to do
Try it. Hold the key, say something, let go. If it's useful, tell someone. If something's wrong, tell me, and I'll fix it.
Elliot