Private dictation for Chrome

Speak messy.
Get clean.

Write the email, capture the thought, finish the draft. Dictaro turns your voice into text and tidies the little things, while your audio stays on your own device.

Chrome 114 or newer · Install the unpacked extension · No account required

From voice to words

One take, four stages

  1. 01
    Speak

    Focus a text field and start from the extension or keyboard shortcut.

  2. 02
    Transcribe

    Whisper turns your voice into text on your own device.

  3. 03
    Clean

    Remove filler, correct your vocabulary and apply spoken editing commands.

  4. 04
    Place

    Cumulative text stays with the field you selected. Your transcript is also ready to copy.

Less typing, more flow

Keep your words. Lose the filler.

Choose Professional, Casual or Verbatim. Add names and phrases to your vocabulary. Say “new paragraph” or “delete last word” to shape the transcript as you speak.

Example spoken input

“um, send the draft to git hub tomorrow”

With vocabulary and cleanup

“Send the draft to GitHub tomorrow.”

Illustrative example with “git hub → GitHub” configured. Review your transcript before sending.

Your voice, your device

Local from microphone to transcript.

Whisper runs inside the extension. Dictaro’s cleanup and vocabulary rules run locally too. Audio and transcripts are not sent to a transcription service.

The first use downloads your selected speech model from Hugging Face. Tiny is approximately 78 MB, Base 148 MB and Small 488 MB. Chrome caches model data in your extension profile; allow enough storage and an internet connection for the first download.

Your model, language, vocabulary and appearance preferences stay in local extension storage. Start and Stop control microphone capture; the popup keeps the finished transcript ready to copy.

Make room for your next thought

Get Dictaro for Chrome.

Download the extension ZIP, extract it and load its folder in Chrome. Choose a model and language, focus a supported top-level text field, and start speaking.

Works with standard text inputs, textareas and editable text in the main page. If a page changes focus or rejects insertion, copy your transcript from the popup.