# kizuna-ai-lab/sokuji

Real-time two-way speech translation for bilingual meetings — auto-detects the spoken language and translates both directions, cloud or fully offline on-device. Desktop (Windows · macOS · Linux) + browser extension (Chrome · Edge) for Zoom, Meet, Teams & any app.

Repository: https://github.com/kizuna-ai-lab/sokuji
Canonical: https://ross.abutalabs.com/products/sokuji
Homepage: https://sokuji.kizuna.ai
Language: TypeScript
License: AGPL-3.0
License Family: copyleft
Topics: ai, meeting, translation, chrome-extension, llm, translator, electron, linux-app, openai, soniox, bilingual, live-translation, real-time, simultaneous-interpretation, speech-translation, stt, tts, two-way-translation
Last push: 2026-09-03T00:30:26+00:00

## Health v2 (maintenance only)
Score: 83/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 100, release rhythm 88, longevity 36
- inputs: {"age_days": 506, "days_push": 0, "days_rel": 1, "gap_med": 1, "n_releases_24m": 166}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1199, forks 171 (observed 2026-09-03T02:15:14.032070+00:00)

## What it is
Sokuji is a cross-platform real-time two-way speech translation app for bilingual meetings, available as a desktop application (Windows, macOS, Linux) and a Chrome/Edge browser extension. It auto-detects the spoken language and translates in both directions, using either cloud providers (OpenAI, Gemini, Soniox, etc.) or fully offline on-device ASR, translation, and TTS powered by WASM and WebGPU.

## Use cases
- translate live speech in real time during bilingual meetings
- interpret Zoom, Google Meet, or Teams calls across two languages
- run speech translation fully offline on-device without an API key
- auto-detect spoken language and translate both directions in a conversation
- add live translated voice output to any video call app
- use a browser extension for simultaneous interpretation in Chrome or Edge

## When to choose
- you need real-time two-way interpretation for bilingual meetings or calls
- you want privacy-focused, fully offline translation without cloud APIs or an expensive GPU
- you want flexibility to switch between multiple cloud providers and local inference
- you need translation that works across Zoom, Meet, Teams, Discord, and any other app

## When to avoid
- you only need text/document translation rather than live speech
- you need broadcast-quality professional simultaneous interpretation for large events
- you require a mobile (iOS/Android) solution — Sokuji targets desktop and browser only

## Facets
- artifact type: application
- maturity: active
- function: speech-recognition, tts, llm-inference, audio-processing, gui
- domain: speech-processing, developer-tools, artificial-intelligence
- platform: windows, browser, wasm, cross-platform
- tags: speech-translation, simultaneous-interpretation, two-way-translation, live-translation, bilingual-meetings, on-device-inference, browser-extension, zoom, google-meet, microsoft-teams, offline-ai, webgpu, natural-language-processing, localization, macos, linux, electron

## Member repositories
- kizuna-ai-lab/sokuji (main) score 83

## Provenance
- Observed fields: from GitHub, fetched 2026-09-03T02:15:14.032070+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:38:02.734565+00:00, confidence not recorded.
  - readme: https://github.com/kizuna-ai-lab/sokuji (fetched 2026-09-03T02:15:14.032070+00:00, sha 5a4622fc14ec)
  - homepage: https://sokuji.kizuna.ai (fetched 2026-08-29T12:42:36.520945+00:00, sha b4bd223e212f)
- Data as of 2026-08-30T08:39:29.467469+00:00.