How to set up Discord live translation locally on Windows

By the end of this guide, incoming Discord voice will appear as live subtitles with real-time AI translation over the Discord desktop app. The app listens and turns speech into captions on your Windows PC. It translates them there too. Your conversations won’t be uploaded to a captioning service.

Key takeaways

• Discord doesn’t include live translation for voice channels.
• OMNI translates the Discord audio playing through your own PC.
• A movable overlay — a caption window shown above another app — puts subtitles directly over Discord.
• Two target languages can appear side by side for every phrase.
• OMNI’s speech recognition and built-in AI translation run on your PC.
• A modern NVIDIA graphics card is recommended for fast captions.

What you need

You only need a Windows PC, Discord, and enough storage for the languages you want to use.

• A 64-bit PC running Windows 10 or Windows 11.
• An NVIDIA graphics card for the best real-time performance.
• OMNI Live Translator, which includes a 30-day free trial without a credit card.
• The Discord desktop app, a Discord account, and access to the voice channel, call, stream, or Stage event you want to follow.
• Headphones or speakers connected to the Windows sound device used by Discord.
• OBS Studio only if you want captions to appear inside a stream. OBS is free streaming software that combines your game, camera, captions, and other items into one picture.
• An internet connection for the first language downloads. OMNI can work offline after setup.

How to set up Discord live translation

The quickest setup is to let OMNI listen to the sound coming from your PC. You can then place its caption window over Discord.

1. Install OMNI and open its language settings. Download the installer from the OmniSubs live page, install it, and find Get more languages. The first launch may take a little longer while the app checks your PC and prepares the available downloads.
2. Download the languages for your conversation. Open Get more languages and download the source and target languages you need. The source is the language being spoken. The target is the language you want to read. Everything runs on your PC after the downloads finish.
3. Choose a performance preset. Start with Balanced. The app will pick the right setup for your PC. Choose Fastest if captions fall behind while you play or stream. Choose Most accurate if your PC can handle more work and exact wording matters most. Don’t judge the result from one shouted sentence over loud music. People speaking over one another and poor microphones make any live captions harder to follow.
4. Set the source and target languages. Pick the language being spoken in Discord as the source. Then choose one or two translation targets. For example, a Japanese gaming stream can show English and Korean together. Picking the correct source language usually helps with names, short callouts, and sentences that mix normal speech with game terms.
5. Match Discord’s output device. In Discord, open User Settings → Voice & Video and check Output Device. This is the speaker or headset Discord uses. Make sure OMNI is listening to the same Windows device in watch-anywhere mode. Join the channel and ask someone to speak. If you can hear Discord but OMNI shows nothing, check these device settings first.
6. Open the native caption window. Turn on the movable caption window and enable its always-stay-on-top setting. Place it over an unused part of Discord or your game. This is the easiest setup for a listener. You don’t need server permissions, a Discord bot, OBS, or control over the speaker’s computer.

7. Join the call, stream, or Stage and test a full sentence. Ask someone to speak normally for several seconds. Testing single words gives you a less useful result. OMNI can wait through a short mid-sentence pause, so it translates the complete thought instead of showing two confusing fragments. On a gaming PC with a modern NVIDIA graphics card, the first captions can appear in under 1 second.
8. Share captions only when your audience needs them. For a casual Discord stream, click Discord’s screen-sharing icon and share the entire screen that contains the caption overlay. If you share only the game window, Discord may leave out the separate caption window. Discord’s current Go Live and Screen Share guide confirms that desktop users can choose an application window or an entire screen. Turn off private notifications before sharing your whole display.
9. Use a browser source for a cleaner stream layout. In OBS, a browser source is a way to show a web page inside your stream. Add a Browser Source, paste OMNI’s browser-source URL, and place the captions over your game or presentation. You can then share the finished OBS view through Discord. Discord can’t use browser-source URLs directly, so OMNI’s movable caption window is still the simpler choice for personal calls and watch parties.

This listener-side setup is especially useful on Discord. You often don’t control the speaker, server, or event settings. OMNI simply translates the sound that reaches your own headset or speakers. A moderator doesn’t need to install anything.

Picking languages: English, Spanish, Russian, Japanese, and Korean

Set the source to the language you expect to hear. Then choose the one or two languages your viewers can read most comfortably. OMNI offers 30+ selectable languages.

For English → Spanish, Balanced is a sensible starting point for normal gaming conversation. Short commands such as “rotate left” are fairly simple. Half-finished jokes spoken by three people at once are not. When possible, ask teammates not to talk over important instructions.

For Russian → English, wait for the complete tactical sentence before reacting. A fragment may leave out details that only become clear at the end. This is why OMNI waits through brief pauses during fast voice chat.

A Japanese → English + Korean setup is useful for game streams, anime watch parties, and community events with mixed audiences. Politeness and the way people address each other carry meaning in Japanese and Korean. Keep the speaker and situation in mind if an English caption sounds unusually blunt. OMNI translates complete sentences and tries to preserve tone, but it can’t know the personal relationship between two speakers.

For English → French + Portuguese, showing two translations lets both community groups follow the same Stage. You don’t need two separate cloud captioning sessions. Try to keep names, abbreviations, and game-specific terms consistent. If you later need polished subtitles for a recording, use OmniSubs’ uploaded-video subtitle workflow instead of treating live captions as final copy.

Guides for other community formats are available for Zoom live translation and Twitch live translation.

Why run translation locally?

OMNI runs on your PC. Discord audio stays there instead of being uploaded to a captioning service. This protects private conversations, removes per-minute charges, and avoids waiting for audio to travel to a cloud service and back. After setup, it can also keep working without an internet connection.

The app listens to speech, turns it into captions, and translates complete sentences. You can choose from 30+ languages. You can also pick Fastest, Balanced, or Most accurate, and OMNI selects the right setup for your computer.

Audio, captions, and translations aren’t uploaded, logged, or stored by an OmniSubs server. That matters in private communities, unreleased game sessions, support groups, and calls where people didn’t agree to send their voices to another cloud provider.

There’s no meter running in the corner either. Per-minute caption bills are a personal pet peeve of mine. They discourage people from leaving accessibility tools on during long, quiet sessions, which is often when everyone forgets they’ll need them.

How does it compare to Discord’s familiar options?

Windows Live Captions is the closest built-in option. Chrome captions and Discord translation bots each solve a smaller part of the problem.

Microsoft’s official Live Captions documentation says normal captioning is available from Windows 11 version 22H2. Translation requires a Copilot+ PC running version 24H2 or later. It’s a strong built-in choice if you have compatible hardware and only need English or Simplified Chinese output.

Chrome supports more captioning options than it once did. It can be useful if you run Discord in a browser. However, it won’t capture sound from the separate Discord desktop app, and translated captions are sent to Google.

A text-translation bot may be enough for a server with multilingual written chat. It can translate messages automatically or after someone enters a command, and everyone can see the result. Voice chat is a different job. Most text bots can’t help when teammates are already talking during a match. Some specialized Discord bots advertise voice features, so check what permissions they need, where they process audio, which languages they support, and how they charge.

FAQ

Does Discord have built-in live voice translation?

No. Discord provides voice channels, calls, streams, Stage channels, text-to-speech, and accessibility settings. It doesn’t currently include a built-in control that captions and translates live voice conversations.

Does OMNI translate what I say back to other people?

No. Watch-anywhere mode translates incoming Discord audio for the person running OMNI. It doesn’t speak a translated version back into the call. Other people can run their own copy or view captions that you include in a shared screen or OBS stream layout.

Which Discord languages can I translate?

OMNI offers 30+ selectable languages, including English, Spanish, Chinese, Hindi, Arabic, Portuguese, French, Russian, Japanese, and Korean. You can display two target languages at the same time.

Are Discord calls uploaded or recorded?

No. OMNI listens, creates captions, and translates them on your PC. It doesn’t upload or log your audio, captions, or translations. After downloading the languages you need, the app can work fully offline.

How much does OMNI cost?

The full app has a 30-day free trial with no credit card. After that, the launch price is a $19.95 one-time lifetime license. There’s no subscription or per-minute fee. The license covers one PC at a time and can be reset or moved up to three times.

Do I need an NVIDIA GPU?

An NVIDIA graphics card is recommended for real-time performance. Start with Fastest or Balanced if Discord, a game, and your streaming software are all using the graphics card at once. The app will choose the right setup for your PC.

Download the OMNI Live Translator free trial to set up local Discord captions and translation on Windows.