freestyle-voice

freestyle-voice / freestyle

Public

Voice-to-text dictation app. Source available, free, and local first.

freestylevoice.com dictation elevenlabs parakeet speech-to-text transcribe
82
11
100% credibility
Found May 28, 2026 at 82 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

Freestyle is a voice-to-text app that lets you dictate anywhere on your computer. Hold a hotkey, speak your words, release, and clean text appears instantly where your cursor is. It connects to AI services you already use, removes filler words automatically, and lets you create custom shortcuts for phrases you say often.

How It Works

1
📥 Download Freestyle

You download the app for your computer and install it in seconds.

2
🔑 Connect your AI service

You enter your API key for a speech-to-text provider like OpenAI or Groq — Freestyle remembers it forever.

3
⌨️ Set your hotkey

You pick a keyboard shortcut like Alt+Space — this is the button you'll hold whenever you want to dictate.

4
🎤 Hold to record

You hold down your hotkey and speak naturally — a floating widget appears showing that you're being heard.

5
✨ Release and watch the magic

You let go of the hotkey and your words appear instantly, cleaned up and polished, exactly where your cursor was waiting.

6
Make it yours
📖
Add shortcuts

Create a dictionary so saying 'my address' automatically becomes your full address.

🧹
Enable cleanup

Turn on AI-powered text polishing to remove 'um' and 'uh' and fix grammar automatically.

📝
Browse history

Review all your past dictations and copy any text you need again.

🎉 Speak 4x faster

You just dictated a full paragraph in seconds — your words, perfectly typed, ready to send.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 82 to 82 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is freestyle?

Freestyle is a cross-platform voice-to-text dictation app that converts your speech into text wherever your cursor sits. Hold down a hotkey, talk, and clean text appears instantly in any application. Built with TypeScript and Electron, it runs locally on macOS, Windows, and Linux, keeping your dictations private since audio never leaves your device. You bring your own API key from providers like OpenAI, Groq, or Deepgram, giving you control over costs and model quality.

Why is it gaining traction?

The "hold to talk, release to paste" workflow is genuinely faster than typing. Unlike browser-based alternatives, this sits as a small floating pill on your screen and works everywhere on your system. The custom dictionary lets you teach it your terminology ("type script" becomes "TypeScript"), and post-processing cleans up filler words automatically. The contextual correction feature is interesting too—it reformats text based on what app you're typing in, like automatically formatting emails properly.

Who should use this?

Developers who write a lot of code comments, documentation, or technical writing will benefit most. It's also useful for anyone who thinks faster than they type. Writers and researchers who need to capture ideas quickly without switching windows will find it valuable. If you're already paying for an OpenAI or similar API key, this is a low-cost way to get professional-grade dictation without subscription lock-in.

Verdict

With 82 stars and a 1.0% credibility score, this is a young project that shows promise but lacks the polish of mature tools. The native binary approach for key injection is clever, but you'll need to grant accessibility permissions and the setup involves some friction. Worth trying if you want local-first dictation with provider flexibility, but check the Discord for support before committing to a workflow.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.