Skip to content

FreeDictate 1.0.0

Choose a tag to compare

View all tags
@DavidHiFi DavidHiFi released this 25 Sep 01:01
· 1 commit to main since this release
v1.0.0
0f42b37

FreeDictate 1.0.0

Unlimited local speech-to-text dictation for Windows 11. Hold Ctrl+Win, speak, release. Text pastes at your cursor. Everything runs on your PC through faster-whisper, so there is no account, no cloud call, and no weekly limit.

Install

Python 3.10 or newer from https://www.python.org/downloads/ first (tick "Add python.exe to PATH"). Not the Microsoft Store build.

Open PowerShell and paste:

irm https://github.com/DavidHiFi/freedictate/releases/latest/download/Install.ps1 | iex

That downloads freedictate.zip, extracts it to %LOCALAPPDATA%\FreeDictate, installs the Python packages, fetches the Whisper model once (about 500 MB), and adds one autostart entry.

Prefer to do it by hand: download freedictate.zip, extract it anywhere, then run

powershell -NoProfile -ExecutionPolicy Bypass -File .\Install-FreeDictate.ps1 -Install

Add -Audit instead of -Install to check first without changing anything.

Use

  1. Click any text field.
  2. Hold Ctrl+Win and speak.
  3. Release. Text pastes at the cursor and stays on the clipboard for a Ctrl+V fallback.
  4. Esc cancels a take with nothing pasted.

One beep when listening starts, a higher beep when text pastes, a low beep when a take is discarded.

If you also use Wispr Flow, quit it or rebind its push-to-talk key first. Both listen to Ctrl+Win, so running both pastes twice.

What it changes

Two things: files in %LOCALAPPDATA%\FreeDictate, and one Run key named FreeDictate under your user hive for autostart. The installer exports that key to backups\ before writing it. It never touches mic routing, audio apps, drivers, display settings, or services.

Uninstall

powershell -NoProfile -ExecutionPolicy Bypass -File .\Install-FreeDictate.ps1 -Uninstall

Then delete the install folder. No restart needed.

Changelog

  • First release. Hold Ctrl+Win push-to-talk, faster-whisper local transcription, auto-paste at cursor, beeps, filler removal, transcript log, one-line installer, read-only audit mode, and one-key uninstall.
  • Added --preload so the installer downloads and caches the model up front instead of on the first dictation.

MIT licensed. Source: https://github.com/DavidHiFi/freedictate

Assets 4
Loading