Why I Built QuickSay: A $49 Dictation App
Voice-to-text tools have been around for years. Dragon NaturallySpeaking was doing it in the 1990s. So why build another one?
Because I got tired of renting one.
The Subscription Problem
I was paying monthly for a voice-to-text tool. It was good. It was also another $15 a month on a pile that already included music, storage, notes, a password manager, and a VPN. Each one seems reasonable on its own. Together they stop being reasonable.
The current generation of voice-to-text tools runs on subscription pricing. Wispr Flow costs $144/year. Aqua Voice runs $96/year. These are useful products, but the pricing model assumes you want a recurring relationship with a voice typing app.
Voice typing is a utility. You press a key, you speak, text appears. A tool like that should be something you buy once and own. So I cancelled the subscription and set out to see whether I could build the same thing myself. It turned out I could.
That’s why QuickSay is $49, one time, for the first 500 customers. $69 for the next 500, then $89 after that. Whichever price you pay, you pay it once, and every future update is included.
Your Key, Both Steps
QuickSay runs on your own free Groq account. You create the key during setup, it gets encrypted on your machine, and from then on your audio goes straight from your computer to Groq. There is no QuickSay server in the middle.
That is unusual in this category. Superwhisper lets you bring your own key for the cleanup step only. Wispr and Aqua don’t allow it at all. QuickSay is the only one where the key is yours for both speech recognition and cleanup.
Two things follow from that. Running costs stay near zero, because Groq’s free tier typically covers it. And there is no middleman who can raise your price later, which is the part that makes “own it forever” mean something.
The honest tradeoff: QuickSay is Windows only. If you’re on a Mac, this isn’t the tool for you, and I’d rather say that plainly than bury it.
The Privacy Question
Some voice-to-text tools offer optional screen capture to add “context” to their transcriptions. I understand the appeal from a product perspective. I wasn’t comfortable building something that can watch your screen while you work.
QuickSay takes a different approach: zero screen capture and zero telemetry. Nothing reads your display, and no usage analytics leave the desktop app. Your audio goes to Groq’s Whisper API for transcription, gets cleaned up by AI, and the text appears in your app. That’s the entire data flow.
How It Actually Works
When you hold the hotkey and speak, QuickSay records audio and sends it to Groq’s Whisper API for transcription. The raw transcript then passes through AI cleanup, which fixes punctuation, removes filler words, and formats the text. The polished result gets placed wherever your cursor is.
The whole process happens in seconds, and it runs on your own Groq key, where the free tier typically covers it.
Staying Small on Purpose
QuickSay’s install size is about 105 MB. Small software tends to be better software: fewer moving parts, fewer things to break, less running in the background.
It isn’t a platform. There’s no AI chat, no screen reading, no workflow automation bolted on. QuickSay turns your voice into text, accurately and fast, and stops there.
The Bet I’m Making
The bet is that there’s a market for tools that are simple, fairly priced, and respectful of privacy. Software doesn’t have to be a subscription to be worth maintaining, and a product doesn’t have to watch your screen to be useful.
QuickSay costs $49 because that’s a fair price for what it does. If it saves you time or makes your work easier, that’s a trade worth making.
Ready to try it? Get QuickSay for $49 and see if voice typing fits into your workflow.