Back to KizunaAI Quick Start

Live Translated Subtitles for Your Talks, Streams and Presentations

Overview

This guide sets up the Sokuji desktop app to turn what you say into translated subtitles for your audience: a talk, a live stream, an online presentation or a lesson. No synthetic voice is generated, and you need no virtual audio cable. The subtitles appear in a floating window that OBS and screen sharing can capture. It uses KizunaAI (Powered by Soniox). The screenshots are from Sokuji 0.41.1 on Windows, speaking 中文 with English subtitles.

1Set up Sokuji with KizunaAI

If you have not done so yet, follow the KizunaAI quick start to install Sokuji and sign in.

2Choose “Subtitle my own speech”

In the setup guide, choose Subtitle my own speech. It puts Sokuji in Me mode with subtitles only. If you have already finished the setup, click Me in the mode switch at the bottom left of the main window and switch on Text Only in Settings.

Choose “Subtitle my own speech”

3Set the two languages

Set I speak to the language you talk in and they read to the language of the subtitles. Here: I speak 中文, they read English.

Set the two languages

4Choose your microphone

In Settings, under Microphone, pick the microphone you speak into.

Choose your microphone

5Try it in the main window

Click Start and say a few sentences. Sokuji shows what you said and its translation as subtitles in the main window. Nothing is spoken aloud.

Try it in the main window

6Open the subtitle window

Click Subtitle in the title bar to turn Sokuji into a slim subtitle window. Move and resize it, then click Start translating. The Both button in its toolbar switches between showing both lines, only the original (Src) or only the translation (Trans).

Open the subtitle window

7Show the subtitles to your audience

In OBS, add a Window Capture source and choose the Sokuji subtitle window. In a video call, share your screen or that window. Your audience sees the subtitles without installing anything.

Tips for Best Results

  • Speak in complete sentences and pause briefly at the end of each: the translation appears once you finish a thought.
  • Show only the translation (Trans) if your audience does not read your language.
  • Use the download button in the toolbar to save the transcript after your talk.

Frequently Asked Questions

Do I need VB-CABLE for this?

No. The virtual audio cable is only needed when others should hear a translated voice. Here your audience reads the translation.

Does my audience need Sokuji?

No. They see the subtitles in your stream, your recording or your shared screen.

Why subtitles only instead of a translated voice?

Your audience keeps hearing your real voice, and generating speech costs extra. If you would rather have a voice, follow the guide for being understood in a meeting.

Troubleshooting

Still stuck? Send us feedback from the account menu, or write to support@kizuna.ai. You can also ask in GitHub Discussions.