Voiceworks Toolkit
A browser-based toolkit for Japanese language learning through immersion with native voiceworks. On-device Whisper transcription, web-assisted translation, and dozens of study features. Built for fast, practical immersion with local caching and WebGPU acceleration when available.
Core Features
Everything you need for immersion-based Japanese study, tuned for speed, clarity, and daily use.
Learner Mode
Real-time dual-language subtitles with Japanese as the primary line and blurrable English translations. Adjustable playback speed and lead-time for study at your own pace.
Live Transcription
On-device Whisper speech-to-text with WebGPU acceleration. Turns any voicework into study material with no manual uploads. Export to LRC, VTT, or SRT formats.
Neural Translation
Web translation pipeline for Japanese and Chinese text with retry logic, host rotation, and shared caching for real-time performance.
Semantic Search
Find voiceworks by meaning using Jina embeddings with local IndexedDB caching. Supports multilingual queries with tag hints and paginated results.
Radio & Playlist Mode
Continuous shuffled playback for extended immersion sessions with auto-recovery. Curated playlists with forward/back navigation and community collections.
30+ QoL Features
Keyboard shortcuts, infinite scroll, audio visualizer, progress tracking, media viewer, interface translation, and much more for a polished study experience.
How It Works
A Tampermonkey userscript that enhances your browsing experience with AI-powered study tools.
Install
Add Tampermonkey to your browser and install the script from Greasyfork with one click. Updates are automatic.
Hybrid AI Pipeline
Whisper transcription runs on-device. Translation and semantic embeddings use web APIs with local caching, retries, and rate-limit handling.
Study Immersively
Browse your library with live subtitles, translations, and keyboard controls designed for hands-free language study.
Companion Tools
Additional utilities for managing and organizing your voicework library.
Voiceworks Renamer
Python desktop application for batch-organizing voicework libraries with metadata enrichment, Whisper transcription (multi-backend: CUDA, ROCm, OpenVINO), audio tagging, and smart playlist generation.
Library Server
Self-hosted media server with Express backend and Vue frontend for streaming your voicework collection. SQLite database, JWT authentication, and filesystem scanning.