VoiceRefine 2.2.1

Fixes

• Fixes the first-run setup crash when no OpenAI API key is configured and the wizard renders Tk canvases under transparent CustomTkinter containers.
• Adds a regression test for concrete Tk canvas backgrounds under transparent CustomTkinter frames.
• Declares customtkinter as an explicit runtime dependency.

Verification

• Python compile check passed for VoiceRefine modules.
• UI background regression tests passed under xvfb.
• Windows PyInstaller build completed on HERM.

SHA256(VoiceRefine.exe): a54e32656ddd41ebd0dc65c64e2220580be31c2b113cbec962ca8fe88be44b61

VoiceRefine 2.2.0 — presets, vault, local Whisper, MCP power-up, UI rewrite

A big release. Read the full notes in CHANGELOG.md.

Highlights

• Named polish presets with per-preset hotkey chords. Win+Alt for default, Win+Alt+C for code, Win+Alt+E for email, Win+Alt+S for summary. Add your own in Settings → Polish.
• First-run wizard — guided onboarding with live API key validation, microphone test (with animated waveform), and press-to-bind hotkey capture.
• Sidebar Settings — seven-section modern UI replacing the old notebook tabs.
• Rich tray menu — last capture preview, pause/resume hotkeys, today's stats.
• New floating overlay — pill-style chip with a 14-bar live waveform.
• Optional Obsidian vault writes — mirror each capture to 11-Data/<category>/YYYY/MM/<id>.json + .md.
• Local Whisper backend (opt-in) — faster-whisper for offline transcription.
• MCP power-up — new tools: `voice_replay_last`, `voice_history_search`, `voice_list_presets`, `voice_polish_to_vault`. Existing tools accept a `preset` argument.
• Press-to-bind hotkey capture everywhere — no more typing `++c` as a string.
• CustomTkinter is the new UI base (raw Tk paths kept as fallback).

Migration

Configs from 2.1.x migrate on first launch — your existing custom prompt becomes your `default` preset, and any custom hotkey is rebound to that preset. Nothing is lost.

Install

Download `VoiceRefine.exe` below, or build from source:

```bash
git clone https://github.com/lordbuffcloud/VoiceRefine.git
cd VoiceRefine
pip install -r requirements.txt
python voicerefine.py
```

Heads up: this build is not yet code-signed. Windows SmartScreen will warn on first launch — click "More info" → "Run anyway". A signed build is coming in a follow-up release.

Wire into Claude Code

```bash
claude mcp add -s user voicerefine python /full/path/to/voicerefine.py --mcp
```

A CK42X project. https://ck42x.com

VoiceRefine v2.1.1

Fixes the hotkey path when Windows has no active microphone input device. VoiceRefine now shows a clear microphone error instead of silently letting the keyboard listener die, adds an input-device selector in Settings, and documents microphone troubleshooting.

VoiceRefine v2.1.0

VoiceRefine v2.1.0 polishes the first-run experience with a guided API key setup, branded translucent Settings and History windows, app/tray icon usage, CK42X links, and a packaged Windows download.