Current build

SpeakFree

A local Windows dictation build designed around one practical moment: press a hotkey, speak, and get text into the app where your cursor already is.

  • Windows-first
  • Local transcription
  • Private by default
  • Current work
SpeakFree product illustration
Building

This page is the portfolio home for the project, with a dedicated download surface now live at speakfree.georgeandary.com.

Download

Try the current Windows build.

The public install path is intentionally simple for this phase: a direct Windows installer, a backup MSI, and a short read on what to expect from the current release.

Primary install

Use the current Windows setup executable if you want the cleanest install path for normal use.

Backup option

The MSI build is also available for anyone who prefers a more traditional installer package.

Current release notes: Windows-first, local dictation, still improving through active QA.

Live now

What the build already does.

The public story here stays grounded in the current app, not in a future version that sounds more polished than the real thing.

Global hotkeys for hold-to-talk and toggle dictation

Local `faster-whisper` transcription with a stronger CPU-first profile

Clipboard-safe paste into the active app

Phrase hints, microphone selection, sound cues, and silence auto-stop

Why it exists

Why this is worth building.

SpeakFree is intentionally narrower than a commercial dictation suite. The aim is a useful daily tool with strong local workflow fundamentals.

Keep dictation local

SpeakFree is being built around privacy and control first, so the core loop does not depend on a cloud transcription product.

Work where the cursor already is

The product is aimed at the practical moment: press a hotkey, speak, and get text into the app you are already using.

Improve through real usage

The build is being tightened against live phrase testing, not just one clean demo clip that makes the system look better than it feels.

Current read

How the project is being framed publicly.

The positioning is simple: useful now, still improving, and grounded in real product iteration instead of startup theater.

Status Building

The current build is usable and improving, but it is still being tightened through active QA.

Platform Windows-first

The shell, hotkeys, and paste behavior are being optimized for Windows 10 and 11.

Positioning Local-first

SpeakFree is intentionally narrower than a polished commercial dictation suite and stronger on privacy and workflow control.

What is next

The next product lift is about reliability, not hype.

The current priority is to keep making dictation feel more natural in real use, especially around insertion reliability, correction memory, and cleaner packaging.

  • Typing fallback for paste-hostile apps
  • A lightweight personal dictionary and correction memory
  • Tray-first idle behavior and cleaner packaging for shareable installs