Skip to content

Releases: bread22/voice-typing

v0.2.0 — Multi-Language & Marketplace Ready

20 Feb 03:25
@bread22 bread22
v0.2.0
604d4ec

Choose a tag to compare

View all tags
v0.2.0 — Multi-Language & Marketplace Ready Latest
Latest

What's New

Multi-Language Support

  • Whisper base model (multilingual) is now the default
  • Auto-detect spoken language (-l auto) — supports Chinese, English, Japanese, Korean, and 90+ languages
  • LLM rewrite preserves the original language of your speech

Extension Icon & Marketplace Polish

  • New AI-themed extension icon (microphone + code brackets)
  • Keywords, categories, and gallery banner for marketplace discoverability

Slimmer Package

  • VSIX reduced from ~634 KB to ~174 KB
  • Excluded source maps and stale build artifacts

Installation

Download voice-prompt-0.2.0.vsix and install:

code --install-extension voice-prompt-0.2.0.vsix

Or in Cursor:

cursor --install-extension voice-prompt-0.2.0.vsix
Assets 3
Loading

v0.1.0 — Cross-Platform Local-First Release

20 Feb 02:46
@bread22 bread22
v0.1.0
4ca116e

Choose a tag to compare

View all tags
v0.1.0 — Cross-Platform Local-First Release

What's New

First public release with full cross-platform support.

Highlights

  • Local-first STT — Uses whisper.cpp CLI directly. No Python, no server, no cloud.
  • Cross-platform audio — Auto-detects SoX (macOS/Linux), arecord (Linux), or FFmpeg (Windows).
  • Auto model download — GGML whisper model (~75 MB) downloads automatically on first use.
  • Works in VS Code and Cursor — Standard VS Code extension APIs only.
  • macOS, Linux, Windows — Platform-specific install instructions in README.

Features

  • Press Alt+V (Option+V on Mac) to start recording
  • Voice Activity Detection (VAD) auto-stops on silence
  • Local LLM rewrite via Ollama (optional) cleans up filler words
  • Cloud rewrite fallback (OpenAI-compatible, opt-in)
  • Cursor stays at end of inserted text with trailing space
  • Configurable model size, VAD sensitivity, rewrite style, and more

Install

  1. Install whisper-cpp: brew install whisper-cpp (macOS) — see README for other platforms
  2. Install an audio recorder: brew install sox (macOS) — auto-detected on all platforms
  3. Download voice-prompt-0.1.0.vsix from this release
  4. In VS Code/Cursor: Ctrl+Shift+P → "Install from VSIX" → select the file
  5. Press Alt+V and start talking!

Optional

  • Install Ollama + ollama pull llama3.2:3b for LLM rewrite
  • Without Ollama, raw transcripts are inserted directly (still useful)
Loading