Skip to content

Releases: liusqu/LocalSpeechtoText_Dialectinput

v2.0.0 — Dialect Voice Input

Choose a tag to compare

@liusqu liusqu released this 11 Apr 06:09

Dialect Voice Input v2.0.0

A macOS menu bar speech-to-text app. 100% local, no cloud, no subscription.

Installation

  1. Download Dialect.Voice.Input.dmg below
  2. Open DMG, drag the app to Applications
  3. First launch: right-click → OpenOpen (Gatekeeper bypass)
  4. Grant Microphone and Accessibility permissions
  5. Wait for model download (~1.7GB, automatic on first launch)
  6. Hold Right Option to talk, release to transcribe

System Requirements

  • macOS 14 (Sonoma) or later
  • Apple Silicon or Intel Mac
  • ~3GB RAM + ~2GB disk

Features

  • Local Qwen3-ASR-0.6B speech recognition (pure C inference engine)
  • 30+ languages, 20+ Chinese dialects
  • Hold / Toggle recording modes
  • Chinese / English UI
  • Auto mutes system audio during recording
  • Optional Cloudflare Tunnel for remote access