Дополнения
Добро пожаловать в дополнения Thunderbird.
Добавляйте дополнительные функции и стили, чтобы настроить Thunderbird по своему вкусу.
Закрыть
Magic Voice Scriber 1.4.3
автор MTF Karukera
Smart semantic voice dictation inspired by Google Rambler: removes filler words, auto-corrects on the fly, and formats clean emails right under your cursor in Thunderbird.
Об этом дополнении
Transform your voice into polished, well-structured emails in Thunderbird 128+ ESR!
Inspired by Google Rambler and Mistral Voxtral, Magic Voice Scriber redefines speech-to-text. Say goodbye to awkward verbatim dictation cluttered with filler words: this extension captures your true spoken intention, eliminates verbal stumbles, and resolves self-corrections on the fly in a single ultra-fast native pass (< 250 ms), with zero latency and no artificial rewriting prompts.
Semantic voice intelligence:
Key highlights:
Inspired by Google Rambler and Mistral Voxtral, Magic Voice Scriber redefines speech-to-text. Say goodbye to awkward verbatim dictation cluttered with filler words: this extension captures your true spoken intention, eliminates verbal stumbles, and resolves self-corrections on the fly in a single ultra-fast native pass (< 250 ms), with zero latency and no artificial rewriting prompts.
Semantic voice intelligence:
Say: "Let's meet on Tuesday... wait no, Wednesday at 2pm to review the draft."
Magic Voice Scriber cleanly inserts:"Let's meet on Wednesday at 2:00 PM to review the draft."
Key highlights:
- Fluent 1-Pass Semantic Dictation (< 250 ms): Speak your mind freely; filler words ("um", "uh"), false starts, and verbal tics are naturally filtered out during audio capture.
- Streamlined Engine Trio (BYOK & Sovereignty):
- Google Gemini (Recommended & Default): Dedicated
gemini-3.5-transcribemodel in native 1-pass SMART mode (immediate semantic cleaning of hesitations). - Mistral AI (High fidelity): Direct 1-pass REST verbatim transcription via Voxtral Mini (
voxtral-mini-latest) for European AI sovereignty. - 100% Offline Local Model: Local Whisper server (Faster-Whisper on
localhost:8000) for total offline privacy and zero cost.
- Google Gemini (Recommended & Default): Dedicated
- In-Situ Voice Caret Gem & Stealth Mode: Discreet micro-trigger under the caret with full stealth mode (strictly hidden at rest when toggled off) and modern animated toggle switches in wizard and options.
- Floating Waveform & Keyboard Controls: Smooth 60 FPS 5-bar reactive sound visualizer powered by Web Audio API, draggable capsule, and complete shortcuts (
Alt+Shift+SorAlt+Shift+D,Spacefor pause,Enterto insert,Escapeto cancel). - 5-Step Onboarding Wizard: Real-time microphone diagnostic, one-click API key testing, live training sandbox, and step-by-step Linux Ubuntu guide (native .deb package).
- Native Multilingual UI: Available in English, French, Spanish, Portuguese, German, and Vietnamese.
