Hedy Multilingual Support in 2026 (36 Languages for Speech, 50 for Chat)
Hedy understands 36 languages for live speech recognition and chats back in 50. See the full lists, the difference between the two, and how to set each in app settings.
Hedy handles two different multilingual jobs: it understands 36 spoken languages in real time, and it chats back to you in 50. This post is the current reference for how that works, which languages are covered on each side, and how to set it up. For the historical milestones, see Hedy’s first multilingual chat update (July 2024) and Hedy’s first multilingual conversation update (October 2024).
Two kinds of language support
Hedy treats the language people speak and the language you read as two separate settings, and that split is the whole point.
- Speech recognition is the language Hedy listens to and transcribes during a session. The transcript, summaries, highlights, action items, and live coaching all come from that spoken language.
- Chat is the language Hedy writes back in. Questions you ask, summaries, recaps, and insights all come out in your chosen chat language, whatever was spoken in the room.
Because the two are independent, a Spanish client call can produce English action items for headquarters, a German engineering meeting can be summarized in Mandarin for an APAC stakeholder, and a Japanese lecture can give you a Portuguese recap.
Languages Hedy understands (speech recognition)
These are the languages Hedy can capture and transcribe live, then analyze for summaries, highlights, action items, and coaching. Speech recognition runs on-device by default, so the audio never has to leave your device. That matters for legal, medical, journalism, and other privacy-sensitive work.
Hedy understands 36 languages for speech recognition:
Afrikaans, Arabic, Catalan, Chinese (Cantonese), Chinese (Mandarin), Croatian, Czech, Danish, Dutch, English, Finnish, French, German, Greek, Hebrew, Hindi, Hungarian, Indonesian, Italian, Japanese, Korean, Malay, Norwegian, Polish, Portuguese, Romanian, Russian, Serbian, Slovenian, Spanish, Swedish, Tagalog/Filipino, Thai, Turkish, Ukrainian, and Vietnamese.
All 36 are available on Mac and Windows. On iPhone, iPad, and Android, Arabic, Hindi, and Thai aren’t yet available for on-device speech recognition, though you can still use them for chat.
Languages Hedy chats in (output)
These are the languages Hedy can respond in. Ask a question and get the answer, summary, or recap in any of them, no matter what language the session was in. Chat reaches more languages than speech recognition because generating written text is an easier problem than recognizing live speech.
Hedy chats in 50 languages:
Afrikaans, Arabic, Bengali, Bulgarian, Burmese, Catalan, Chinese (Cantonese), Chinese (Simplified), Chinese (Traditional), Croatian, Czech, Danish, Dutch, English (UK), English (US), Estonian, Farsi, Finnish, French, German, Greek, Hebrew, Hindi, Hungarian, Indonesian, Italian, Japanese, Korean, Lithuanian, Macedonian, Malay, Malayalam, Norwegian, Polish, Portuguese (Brazilian), Portuguese (European), Romanian, Russian, Serbian, Sinhala, Slovak, Slovenian, Spanish, Swedish, Tagalog/Filipino, Tamil, Thai, Turkish, Ukrainian, and Vietnamese.
How to set it up
- Open Hedy and go to Settings → Speech & AI.
- Pick your speech language, or use auto-detect if your meetings switch languages.
- Pick your chat language separately if you want Hedy to respond in a different language than the one spoken.
- Start a session. Hedy transcribes in the spoken language and responds in your chat language.
You can also set the language per session when you create it, so you don’t have to change global settings every time you switch contexts.
Who benefits most
- International teams holding cross-border meetings where participants prefer different languages
- Students and academics in multilingual classes, lectures, and conferences
- Coaches and consultants with clients in different markets who want notes in the client’s language and reports in their own
- Journalists interviewing sources in one language and writing copy in another
- Anyone learning a language who wants to follow along in the target language while Hedy explains things in their native one
On-device, cloud, and data residency
Speech recognition runs on-device by default, which keeps audio off any server. Optional cloud providers (Deepgram, OpenAI) can improve accuracy on specific languages, but require sending audio off-device. Both options live in Settings → Speech & AI.
For users who select EU data residency, all conversation processing, including multilingual analysis, runs through European infrastructure.
“Every conversation is an opportunity to shine, regardless of what language you’re most comfortable in,” says Julian Pscheid, founder of Hedy AI. “We want more people, in more languages, to access Hedy’s full capability in a way that feels natural.”
Looking ahead
We keep adding languages as model quality improves and as users tell us what they need. The language pickers in app settings are always the current authoritative source. If a language you need isn’t there yet, let us know. User demand is what drives the roadmap.
About the author
Julian Pscheid is the founder and CEO of Hedy AI, a real-time AI meeting coach used by tens of thousands of professionals worldwide. He writes about how AI is changing the way people prepare for, capture, and understand important conversations.
