yebisu0529-ship-it/talk-style-live-translation

How yebisu0529-ship-it/talk-style-live-translation works

Talk風Live翻訳 is a Windows desktop app that translates PC audio into Japanese in a chat-style transcript, with local translation and speech recognition. The principal workflow is: the user starts listening, the app obtains audio or another selected input, speech is recognized, text is translated, and results appear in the chat and history. Separate branches provide text, document, screen/OCR, subtitle, and read-aloud features, alongside model setup. The graph reflects README-described capabilities and sampled wiring; several implementation paths are unsampled, so uncertain connections are omitted.

Main entry points: main.rs App orchestration · chat.rs Chat transcript · setup_chain.rs Setup sequence · setup.rs First-run setup

Components 26

Desktop app

App UI src/ui.rs
App orchestration src/main.rs
Chat transcript src/chat.rs
Talk history src/journal.rs

Live translation

PC audio input src/platform.rs
Screen feed src/screen_feed.rs
YouTube captions extension/background.js
Speech recognition src/nemotron.rs
ASR service scripts/asr_server.py
Translation client src/engine.rs
Local translation runtime

Translation features

Simple translation src/simple.rs
Document translation src/document.rs
Screen and image OCR src/screen_ocr.rs
Subtitle tools src/subtitles.rs
Read aloud src/readout.rs

Models and setup

Model selection src/model_catalog.rs
Hardware assessment src/hardware.rs
First-run setup src/setup.rs
Setup sequence src/setup_chain.rs
Model installer src/installer.rs
Speech preparation src/speech_setup.rs

Speech and display

Subtitle window src/subtitle_window.rs
Theme and styling src/theme.rs
Voice profiles src/voice_profile.rs

Other components

User

Connections 26

User → App UI: uses README.md
App UI → App orchestration: controls src/main.rs
App orchestration → PC audio input: captures audio src/main.rs
App orchestration → Screen feed: starts screen feed src/main.rs
App orchestration → YouTube captions: accepts captions
App orchestration → Speech recognition: starts recognition src/main.rs
Speech recognition → ASR service: sends audio
App orchestration → Translation client: requests translation src/main.rs
Translation client → Local translation runtime: calls local API src/engine.rs
App orchestration → Chat transcript: updates transcript src/main.rs
Chat transcript → Talk history: records talks
App orchestration → Simple translation: opens feature src/main.rs
App orchestration → Document translation: opens feature src/main.rs
App orchestration → Screen and image OCR: opens feature src/main.rs
App orchestration → Subtitle tools: opens feature
App orchestration → Read aloud: polls readout src/main.rs
App orchestration → Subtitle window: updates window src/main.rs
First-run setup → Model selection: selects model
Model selection → Hardware assessment: checks fit src/model_catalog.rs
First-run setup → Setup sequence: plans preparation src/setup.rs
Setup sequence → Model installer: installs components
Setup sequence → Speech preparation: prepares speech
App orchestration → First-run setup: runs setup src/main.rs
Subtitle window → Chat transcript: renders messages src/subtitle_window.rs
Chat transcript → Theme and styling: uses styling src/chat.rs
Read aloud → Voice profiles: uses voice settings

Drawn by GitDiagram on November 14, 2026 from the repository's file tree, README and source files. Read as Markdown ·