Windows-native • Local-first • No account required

Talk to your computer. Shruti does the typing.

Trigger dictation from anywhere and send the final text back into your focused app automatically — powered by local speech recognition.

100%
local transcription
Ctrl+Win+Space
default hold-to-dictate
3
insertion modes
Shruti home screen showing dictation status
Local Whisper transcription Clipboard fallback Tray and floating controls Light and dark themes

Designed for daily writing

Dictation that stays out of your way.

Shruti focuses on one job: quickly turning speech into useful text inside messages, prompts, notes, and documents.

Global dictation shortcut

Start from anywhere with the default Ctrl+Win+Space trigger, tray menu, or compact floating microphone.

Insert into your active app

Shruti remembers your previous target, restores focus, and inserts the final transcript with a clipboard fallback when needed.

Live preview, final output

See partial text while recording. Only the finalized transcript is inserted, so your work stays clean.

Local model management

Download, verify, import, and remove speech models without creating an account or sending recordings to a hosted service.

Auto, preview, or copy-only

Choose immediate insertion by default, review before insert, or keep dictation as a copy-to-clipboard workflow.

Native Windows feel

Built with WinUI 3, a tray presence, system theme support, and microphone controls that feel at home on Windows.

Privacy by default

Your voice stays on your machine.

Shruti is local-first. Core transcription runs on your PC, and models, settings, recordings, and transcripts live locally unless you explicitly export them.

  • No hosted transcription required for the core workflow
  • No account or subscription gate planned for local dictation
  • Transcript text is redacted from diagnostic logs by default
Shruti settings screen for model and app preferences

Simple loop

From voice to text in five steps.

01

Trigger

Use the shortcut, tray, app button, or floating mic.

02

Capture

Shruti remembers the foreground app and records your microphone.

03

Transcribe

Audio is normalized and processed locally through the selected model.

04

Review

Watch live partial text or open preview-before-insert mode.

05

Insert

The final transcript lands back in the app you were using.

Open source

Follow or contribute to Shruti.

Shruti is under active development in the open. Explore the source, build the Windows app locally, or follow the project for release updates.