Skip to content

The Live AI Translation Hub

Let everyone be part of the moment.

One platform. Live AI translation forstages., sessions., webinars., performances., films., TV., live broadcasts., recordings., conversations., calls., gaming.

Request a pilot

Got a code from your event?

Live demo

Join as a listener.Pick any language.

The real listener page with real Tõlk translations. Choose a broadcast, a language and how you listen. Captions, AI Summary, agenda and speakers included.

118 languages. Each listener picks theirs.

Speakers can talk in 118 languages, and listeners hear translations in 81, across every model Tõlk runs. Every name on the globe is a language Tõlk translates into, written the way its speakers write it. Give it a spin.

Speech in
118
Speech out
81
  • Afrikaans
  • Akan
  • Shqip
  • አማርኛ
  • العربية
  • Հայերեն
  • Azərbaycan dili
  • Euskara
  • Беларуская
  • বাংলা
  • Български
  • မြန်မာစာ
  • Català
  • 中文
  • 简体中文
  • 繁體中文
  • Hrvatski
  • Čeština
  • Dansk
  • Nederlands
  • English
  • Eesti
  • Filipino
  • Suomi
  • Français
  • Galego
  • ქართული
  • Deutsch
  • Ελληνικά
  • ગુજરાતી
  • Hausa
  • עברית
  • हिन्दी
  • Magyar
  • Íslenska
  • Bahasa Indonesia
  • Italiano
  • 日本語
  • Basa Jawa
  • ಕನ್ನಡ
  • Қазақ тілі
  • ខ្មែរ
  • Ikinyarwanda
  • 한국어
  • ລາວ
  • Latviešu
  • Lietuvių
  • Македонски
  • Bahasa Melayu
  • മലയാളം
  • मराठी
  • Монгол
  • नेपाली
  • Norsk
  • Norsk bokmål
  • فارسی
  • Polski
  • Português
  • Português (Brasil)
  • Português (Portugal)
  • ਪੰਜਾਬੀ
  • Română
  • Русский
  • Српски
  • سنڌي
  • සිංහල
  • Slovenčina
  • Slovenščina
  • Español
  • Basa Sunda
  • Kiswahili
  • Svenska
  • தமிழ்
  • తెలుగు
  • ไทย
  • Türkçe
  • Українська
  • اردو
  • Oʻzbekcha
  • Tiếng Việt
  • isiZulu

Different moments. One hub.

Five products. Same languages, engines and delivery underneath.

Compare products

Starting prices. Up front.

Per language, per hour. Listeners never cost extra. Your quote comes from our team.

Tõlk at a real event.

Real footage: live captions on a laptop, a listener on their phone, the AV desk on air. Turn the sound on.

0.4 sfrom speech to translated audio
Recorded at a live event: captions, phones, and the AV desk.

AI Summary

Joined late? Catch up in your language.

Every few minutes Tõlk writes down what was said, in each listener’s language. After the event, the recap goes out by email.

  • First summary about 5 minutes in, then every 5 minutes
  • In each listener’s own language
  • Last 5 minutes, earlier in the session, key quotes
  • Recap and transcript by email after the event

Summaries are AI-generated and may contain mistakes. The organizer decides when the recap is shared.

AI SummaryEnglish
Illustration of the listener view.

Speaker voices

Every speaker, their own voice.

On a panel, listeners can tell who is talking by the voice alone. Each speaker keeps one voice, whichever language you listen in.

  • Pick a voice for each speaker in the program
  • Or let pronouns choose: he or she, else the language’s voice
  • The same voice in every language
  • Captions name who is talking

For languages read by a studio voice or Dictionary-assisted. Direct speech-to-speech models speak with their own voice.

Main stageEspañol
EspañolLiis Saar’s voice
Illustration of the listener view. Names are examples.

Raise a hand

Ask in your language. Keep talking.

Anyone in the audience can raise a hand on their phone and speak in their own language. Speakers hear it in theirs, answer, and the conversation carries on without a barrier.

  • Raise a hand from the join page, no app
  • Speak in your own language through your phone
  • Each speaker hears it in theirs, with captions
  • Approve each turn, or let hands go live straight away

Up to 8 open mics at once. A dedicated audience mic in the room works too.

Main stageEesti
Illustration of the listener view. Names are examples.

Under the hood

The best mix of models. None of the setup.

Leading speech-to-speech models and speech → text → speech cascades in one platform. No provider accounts, no API keys.

For AV teams and developers

Built for the AV desk. And for developers.

Every input your production has. Every output it needs. An API for the rest.

  • Every way in
  • Multi-mic mixer
  • AV desk view
  • Every way out
  • One-click restream
See the AV and developer tools

Connect Tõlk to Zoom, Microsoft Teams, YouTube Live, OBS Studio, vMvMix, ClickMeeting, Twitch, and anything that speaks RTMP.

Take the audio in from your meeting, stream or mixer. Send each language back out. These work through each platform’s own streaming features: workflows, not partnerships.

See all integrations
How Tõlk works Routes and models in detail

Real situations. Pick yours.

Photos are illustrations, not customer deployments.

Conference audience facing a stage ConferenceDelegates at a conferenceThe keynote in their language, by QR code.Visitors in a gallery SpotlightA guide with 30 visitorsEach visitor hears the guide in their language.Broadcast control room StudioA team with recorded coursesReviewed voices and captions per language.Two people talking at a table PolylogueColleagues without a shared languageEveryone, in their own language. BridgeA gamer in a cross-language guildTeammates hear them in Spanish on Discord.Engineer at a mixing console ConferenceAn AV engineerOne feed in. A stream per language out.
All solutions

In the field

Built at real events.Not in a lab.

Live events and multi-day seminars, several stages, many languages. Case studies come with the organizers’ permission.

Hear the demosStories

Need a custom workflow?

Venues, agencies, media teams. We fit Tõlk to how you already work.

Custom solutions

Building with translation?

Start sessions from your platform. Get listener and stream links back.

API documentationFor developers

The Live AI Translation Hub

Ready for your next moment?

Tell us what to make multilingual. We pilot it on your content, then you decide.

Request a pilotTalk to the team · hello@tolk.fm