There's something you've known how to do since you were two: talk.
And yet today, if you want your words to become text, the market has decided it's a luxury. A monthly subscription to dictate an email. A “Pro” plan to remove the weekly word cap. Another subscription, per seat, to get a summary of a meeting. And while you pay, your voice travels: it leaves your computer, crosses servers you've never heard of, and somewhere along the way becomes what the contracts call “data”.
It wasn't always like this. The promise of personal software used to be different: you need a tool, you get it, it's yours. Today the promise has changed: you need a tool, so we hold you hostage. A little every month, forever.
Meanwhile, across the ocean, the numbers now look like this: dictation startups that have raised more than eighty million dollars and negotiate at valuations whispered in billions. Meeting-notes apps valued at a billion and a half. They're good, some of them are excellent. But those millions have to be paid back, and there's only one place they can come from: you. Your subscription that goes up, your free plan that shrinks, your data working for someone else. They're not bad people. It's the model that works this way: you're not the customer to serve, you're the metric to grow.
I lived this every day. I'm a developer, I live in Bologna, and I dictate for a living: emails, code, notes, messages. At some point I did the math: how much I was paying per year in subscriptions to turn voice into text. And meanwhile I could see the other half of the picture: local voice models have gotten really good. What needed a data center three years ago now runs on the laptop in front of you, locally, in under a second. The math didn't add up. Not technically, and not as a principle.
So I did the most naive thing in the world: I decided to try building it myself.
No investors. No millions raised. Nights, weekends, and a group of friends who tried every version, broke every build, and told me without mercy what was bad. They don't appear on any org chart, because there is no org chart. But without them Dimmy wouldn't exist.
We called it Dimmy because it sounds like “dimmi”, Italian for “tell me”. That's exactly what it does: you tell me, I'll handle the rest.
And since we had no investors to pay back, we could do something rare: always choose the side of the person using it. Every technical choice is a choice of sides.
Your voice stays on your computer. Transcription runs locally, audio is never written to disk and never leaves your machine, unless you choose a cloud provider. And that's the point: everything that leaves your machine is your decision. What to enable, how, when. The user chooses. Not the software house, not the giants, not the people funding them. We don't ask for an account to get started. There's no server of ours collecting what you say, because there's no server of ours at all.
The code is public. All of it, under a free license. If you can build it yourself, Dimmy is yours, free, forever, with nothing crippled. If you pay, you're not buying features: you're buying the convenience of a signed build, automatic updates, and a human who answers. It seems like the most honest deal possible.
And the price is a price, not an annuity: a few euros a month, or a one-time amount, no renewals. Set by looking at what the work is worth, not at what the market will bear.
The result is a small, fast tool that does three things. You dictate anywhere, in 99 languages, and the text appears where you're writing, already clean and in the style you decide, even automatically per app: formal in email, casual in chat. You select text, speak, and transform it with a command. And when you have a meeting, it transcribes it and hands you the recap with decisions and action items, without sending any bot into the call. The audio stays on your computer; the recap is generated locally or with the provider you choose.
That's it. You speak, text appears. Under a second. It seems like a small thing, until you notice how much they were charging you for it.
Now, let's be honest: out there are giants with hundreds of people and nine-figure bank accounts. Not us. But that's exactly why we answer to one person only: you.
We want your voice to be heard. Better: we want to give it a superpower.
That's not a slogan, it's literal. Your voice is the most natural tool you have, and in front of a computer it has been mute for years. With Dimmy you speak, and text appears. You speak, and the meeting sums itself up. You speak, and things get done.
Making software this way, local, with no investors, at an honest price, is a consequence for us, not a flag. It's how we started, and it felt like a good road from day one. The rest is math you can do on your own: a subscription every month, forever, or a tool that's yours. Your voice on someone else's servers, or your voice staying yours.
We imagine a small, reasonable world: where simple things stay simple, where your voice counts and isn't counted, and where talking to your computer doesn't require a lifelong subscription.
Dimmy is our piece of that world.
Dimmi. Tell me. We'll handle the rest.
Konrad, Bologna