Understand spoken conversations
Capture tab or microphone audio for bilingual captions. Show one or four caption groups and customize fonts, colors, and outlines.
VocaLumeA little closer, in any language.LOCAL TRANSLATION · PUBLIC BETA
The jokes in a stream. The little moments in chat. A thank-you you want to share.
Follow the conversation, and reply in your own language.
No paid API needed in local mode. Download models on first use.
Explore the site and demos on your phone. Install the translation app and extension on a computer.
Preset example, not a live translation service
Japanese speech becomes Chinese captions.
Follow the stream and read the chat.
A 20-second scenario using the actual caption colors.
Recorded captions, simulated UI.
#まほろくりっぷ · Unofficial; no affiliation or endorsement.
Read your friend’s Japanese, then reply in your own language.
Choose “ZH → JA”, check the draft, and press Enter to send.
Simulated chat. Review translations before sending.
Viewers see Traditional Chinese captions for Japanese speech in OBS. The creator reads Japanese translations of foreign-language chat in a separate panel.
Synthesized Japanese speech; translated chat is visible only to the streamer.
Choose separate languages for reading, typing, and speech.
Or link reading and reply directions for a simpler setup.
Capture tab or microphone audio for bilingual captions. Show one or four caption groups and customize fonts, colors, and outlines.
Translate text on supported websites. You decide when to enable it, so translations only appear when you want them.
Type in Chinese and get a Japanese draft. Check the wording and tone, then send it yourself. Other supported language pairs are available too.
Use the desktop app to translate your microphone audio and send captions to OBS. The browser extension translates foreign-language chat on YouTube or Twitch for the streamer.
Get the app for your device from GitHub, follow the extension setup guide, and download the models.
Choose the spoken and translation languages. For chat, you can link reading and reply languages in opposite directions.
Play a video and start captions. To reply, translate your draft, review it, and send it yourself.
4.6.6 includes Windows, Mac and experimental Linux apps plus the Chrome / Opera GX extension. Local Mac browser captions require one-time app registration; follow the installation guide.
Choose your microphone and translation language in the VocaLume desktop app, then enable OBS caption output. Add a text source in OBS, enable Read from file, and select the caption file shown in the app. Include that text source in your stream scene so viewers can see it.
Success and accuracy are not guaranteed. Audio quality, background music, overlapping speakers, accents, sentence boundaries, names, device performance, and the selected model can affect results. Short phrases without context may be mistranslated. If a translation fails, follow the interface prompts to retry or check caption history. Verify important content against the original. We do not yet have success-rate or accuracy figures representative of all devices, languages, and situations.
Local speech recognition and translation run on your computer without a paid API. Models download on first use. Optional external APIs are billed by their providers.
Speed depends on your hardware, model, audio, and system load. Try the public beta to see how it performs on your device. Local processing uses battery and hardware resources.
Website compatibility is still improving, and some sites or editors have limitations. Follow the setup guide and report issues on GitHub with your environment details.
No. Translations become drafts for you to review and send. Recognition and translation can make mistakes, so double-check important messages.
PUBLIC BETA · STILL EVOLVING
VocaLume is not a final release. Translation quality, performance,
and website compatibility will continue to improve with future updates.
Follow the project’s progress, or tell us what could work better for you.
A Star bookmarks the repository. For release notifications, use GitHub Watch → Custom → Releases.