Hold a key. Speak. Release. Dictio turns what you said into clean, punctuated text and types it right where your cursor is, in any app, on any of your devices.
Press and hold the hotkey. A small waveform appears so you know it's listening.
Say what you mean. Emails, prompts, tickets, messages, notes. Any of 50+ languages.
Your words land in whatever text field you were in, cleaned up and ready to send.
Dictio isn't a separate window you dictate into and copy from. It types into the app you're already using, on Mac, Windows, iPhone, and Android. If the cursor blinks there, Dictio can write there. And everything is copied to your clipboard too, so nothing is ever lost.
The same sentence, typed versus dictated. Timed, not estimated.
Use Dictio with Claude Code, Cursor, Copilot, or any tool with a text field. No plugin, no integration, no touching your keyboard.
Terminal, IDE, browser, chat. If it takes text, Dictio pastes into it. No special integration required.
Describe the architecture, the edge case, the constraint. Out loud, at around 150 words per minute, while the cleanup strips the filler words for you.
metering.py can you add a daily spend cap check before the transcribe call. Read the limit from env and return a clear error if we're over it, not a silent failure like it is now.
Dictio works with your own OpenAI API key. Your audio goes straight from your device to OpenAI under your key, and it never passes through our servers. We can't hear you, we can't store you, and we have nothing to train on.
Stored where your passwords live. Never written to a plain file.
Recordings and transcripts stay local, for your eyes only. Delete them whenever you like.
The questions everyone asks before their first dictation.
Mac, Windows, iPhone, and Android. One purchase covers your devices.
No. $19 is the whole price for the app. Transcription runs on your own OpenAI key, and you pay OpenAI only for what you actually use, which for most people is a few cents a day.
Straight from your device to OpenAI, under your own API key. It never passes through Zehntech servers, and we have no way to hear, store, or train on anything you say. Full details on the privacy page.
No account, no sign-up. Install the app, add your OpenAI key, and start dictating. Creating a key takes about two minutes, and we show you how right after install.