fayazara

fayazara / Kaze

Public

Local speech to text app - Talk instead of typing.

89
4
100% credibility
Found Feb 25, 2026 at 53 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Swift
AI Summary

Kaze is a macOS menu bar utility that captures speech via a global hotkey, transcribes it using local engines, and pastes the text into the active application.

How It Works

1
🔍 Discover Kaze

You find Kaze, a simple Mac tool that turns your speech into text and pastes it anywhere with a quick key hold.

2
💻 Get it running

Follow the easy steps to install Kaze on your Mac, and it quietly sits in your menu bar ready to go.

3
🔑 Allow access

Give Kaze permission to hear your voice and watch for your special key combo in your Mac's privacy settings.

4
Pick speech style
🍎
Built-in dictation

Uses your Mac's ready-to-go voice recognition, no extra steps.

🤖
Advanced listener

Grab a tiny download once for offline, high-quality speech capture.

5
✨ Tweak text polish

Optionally turn on smart cleanup to fix grammar and punctuation automatically.

6
🎤 Hold and speak

Press and hold Option-Command, talk naturally while watching the fun floating waveform show your voice levels and growing text.

✅ Text pastes instantly

Release the keys, and your polished words appear right in whatever app you were using, clipboard safe.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 53 to 89 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 Kaze?

Kaze is a Swift-based macOS menu bar app that turns speech into pasted text anywhere: hold Option+Command to record, speak, and release to auto-paste the transcription into your focused app like Xcode, Notes, or a browser. It ditches typing drudgery with fully local processing via Apple's dictation or Whisper models, plus a subtle bottom-screen waveform overlay for live feedback. Clipboard stays safe—original contents restore post-paste.

Why is it gaining traction?

Push-to-talk hotkeys work globally without app-switching, dual engines offer zero-setup dictation or offline Whisper accuracy, and Apple Intelligence optionally refines grammar on-device. No dock clutter, just settings for engine choice and custom prompts; it's a privacy-first local github copilot alternative to cloud STT like kaze ai github tools. At 45 stars, it hooks macOS devs wanting emanuar-style local github runner simplicity for voice.

Who should use this?

MacOS devs dictating API docs, commit messages, or Slack replies mid-coding flow. Designers sketching wireframe notes in Figma or Sketch. Podcasters or writers transcribing ideas into Ulysses without cloud risks, especially on Apple Silicon.

Verdict

Promising for local speech-to-text but immature: 45 stars, 1.0% credibility score, good README, no tests. Grab if you're on macOS 15+ and need a kazenmaier leasing-free local github copilot alternative—fork and contribute to mature it.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.