Skip to main content
WhisperAI
Powered byOpenAI
Cloud SyncWhisper API
  1. Home
  2. Supported Languages
Language coverage

Every language WhisperAI supports.

WhisperAI transcribes 98 languages from audio and video, streams 18 of them live, and translates a finished transcript into 87 target languages. The table below is generated from the same lists the product and the Developer API use, so the codes here are exactly the codes an API request accepts.

98 transcription
18 realtime
16 multi-language
87 translation targets

Four lists, not one

“Supported” means something different for each capability, and the lists genuinely differ. A language can be transcribable but not offered as a translation target, or available for uploaded files but not for live streaming.

98 transcription languages

Upload audio or video and get text back. 101 request codes in total, because English is offered in four regional variants. Leave the language on auto-detect and WhisperAI picks it per file.

18 realtime languages

Live streaming transcription runs on a smaller, curated set than file transcription. Anything outside it should be transcribed as a file.

16 multi-language languages

Multi-language (code-switching) mode handles recordings where speakers move between languages mid-sentence, across this subset.

87 translation targets

Translate a finished transcript into other languages — 91 target codes once regional variants of English and Chinese are counted separately.

The full list

Search by language name or ISO code, or filter by the capability you need. Languages with a dedicated guide link through to it.

Filter by capability

Showing 99 of 99 languages.

Languages supported by WhisperAI, with ISO codes and per-capability availability
LanguageISO codeTranscriptionRealtimeMulti-languageTranslation
A
Afrikaansaf●Supported–Not supported–Not supported●Supported
Albaniansq–Not supported–Not supported–Not supported●Supported
Amharicam●Supported–Not supported–Not supported●Supported
Arabicar●Supported●Supported–Not supported●Supported
Armenianhy●Supported–Not supported–Not supported●Supported
Assameseas●Supported–Not supported–Not supported●Supported
Azerbaijaniaz●Supported–Not supported–Not supported●Supported
B
Bashkirba●Supported–Not supported–Not supported–Not supported
Basqueeu●Supported–Not supported–Not supported●Supported
Belarusianbe●Supported–Not supported–Not supported●Supported
Bengalibn●Supported–Not supported–Not supported●Supported
Bosnianbs●Supported–Not supported–Not supported●Supported
Bretonbr●Supported–Not supported–Not supported–Not supported
Bulgarianbg●Supported–Not supported–Not supported●Supported
Burmesemy●Supported–Not supported–Not supported–Not supported
C
Catalanca●Supported–Not supported–Not supported●Supported
Chinesezh, zh-CN, zh-TW●Supported●Supported●Supported●Supported
Croatianhr●Supported–Not supported–Not supported●Supported
Czechcs●Supported–Not supported–Not supported●Supported
D
Danishda●Supported●Supported–Not supported●Supported
Dutchnl●Supported●Supported●Supported●Supported
E
Englishen, en_au, en_uk, en_us●Supported●Supported–Not supported●Supported
Estonianet●Supported–Not supported–Not supported●Supported
F
Faroesefo●Supported–Not supported–Not supported–Not supported
Finnishfi●Supported●Supported●Supported●Supported
Frenchfr●Supported●Supported●Supported●Supported
G
Galiciangl●Supported–Not supported–Not supported●Supported
Georgianka●Supported–Not supported–Not supported●Supported
Germande●Supported●Supported●Supported●Supported
Greekel●Supported–Not supported–Not supported●Supported
Gujaratigu●Supported–Not supported–Not supported●Supported
H
Haitianht●Supported–Not supported–Not supported●Supported
Hausaha●Supported–Not supported–Not supported●Supported
Hawaiianhaw●Supported–Not supported–Not supported●Supported
Hebrewhe●Supported●Supported–Not supported●Supported
Hindihi●Supported●Supported●Supported●Supported
Hungarianhu●Supported–Not supported–Not supported●Supported
I
Icelandicis●Supported–Not supported–Not supported●Supported
Indonesianid●Supported–Not supported–Not supported●Supported
Italianit●Supported●Supported●Supported●Supported
J
Japaneseja●Supported●Supported●Supported●Supported
Javanesejw●Supported–Not supported–Not supported●Supported
K
Kannadakn●Supported–Not supported–Not supported●Supported
Kazakhkk●Supported–Not supported–Not supported●Supported
Khmerkm●Supported–Not supported–Not supported–Not supported
Koreanko●Supported–Not supported●Supported●Supported
L
Laolo●Supported–Not supported–Not supported●Supported
Latinla●Supported–Not supported–Not supported●Supported
Latvianlv●Supported–Not supported–Not supported●Supported
Lingalaln●Supported–Not supported–Not supported–Not supported
Lithuanianlt●Supported–Not supported–Not supported●Supported
Luxembourgishlb●Supported–Not supported–Not supported●Supported
M
Macedonianmk●Supported–Not supported–Not supported●Supported
Malagasymg●Supported–Not supported–Not supported●Supported
Malayms●Supported–Not supported–Not supported●Supported
Malayalamml●Supported–Not supported–Not supported●Supported
Maltesemt●Supported–Not supported–Not supported●Supported
Maorimi●Supported–Not supported–Not supported●Supported
Marathimr●Supported–Not supported–Not supported●Supported
Mongolianmn●Supported–Not supported–Not supported●Supported
N
Nepaline●Supported–Not supported–Not supported●Supported
Norwegianno●Supported●Supported–Not supported●Supported
Norwegian Nynorsknn●Supported–Not supported–Not supported–Not supported
O
Occitanoc●Supported–Not supported–Not supported–Not supported
P
Panjabipa●Supported–Not supported–Not supported●Supported
Pashtops●Supported–Not supported–Not supported●Supported
Persianfa●Supported–Not supported–Not supported●Supported
Polishpl●Supported–Not supported●Supported●Supported
Portuguesept●Supported●Supported●Supported●Supported
R
Romanianro●Supported–Not supported–Not supported●Supported
Russianru●Supported–Not supported●Supported●Supported
S
Sanskritsa●Supported–Not supported–Not supported–Not supported
Serbiansr●Supported–Not supported–Not supported●Supported
Shonasn●Supported–Not supported–Not supported●Supported
Sindhisd●Supported–Not supported–Not supported●Supported
Sinhalasi●Supported–Not supported–Not supported●Supported
Slovaksk●Supported–Not supported–Not supported●Supported
Sloveniansl●Supported–Not supported–Not supported●Supported
Somaliso●Supported–Not supported–Not supported●Supported
Spanishes●Supported●Supported●Supported●Supported
Sundanesesu●Supported–Not supported–Not supported●Supported
Swahilisw●Supported–Not supported–Not supported●Supported
Swedishsv●Supported●Supported–Not supported●Supported
T
Tagalogtl●Supported–Not supported–Not supported●Supported
Tajiktg●Supported–Not supported–Not supported●Supported
Tamilta●Supported–Not supported–Not supported●Supported
Tatartt●Supported–Not supported–Not supported–Not supported
Telugute●Supported–Not supported–Not supported●Supported
Thaith●Supported–Not supported–Not supported●Supported
Tibetanbo●Supported–Not supported–Not supported–Not supported
Turkishtr●Supported●Supported●Supported●Supported
Turkmentk●Supported–Not supported–Not supported–Not supported
U
Ukrainianuk●Supported–Not supported●Supported●Supported
Urduur●Supported–Not supported–Not supported●Supported
Uzbekuz●Supported–Not supported–Not supported●Supported
V
Vietnamesevi●Supported●Supported●Supported●Supported
W
Welshcy●Supported–Not supported–Not supported●Supported
Y
Yiddishyi●Supported–Not supported–Not supported●Supported
Yorubayo●Supported–Not supported–Not supported●Supported

English is available as four regional codes — en (Global), en_us, en_uk and en_au — and Chinese has separate Simplified (zh-CN) and Traditional (zh-TW) translation targets. That is why the code counts (101 transcription, 91 translation) are higher than the language counts.

How language selection works

Auto-detect is the default

Leave the language unset and WhisperAI detects it per file. Set it explicitly when you already know the language and the audio is short, noisy or heavily accented — pinning the language stops detection from drifting mid-recording.

Mixed-language recordings have their own mode

If speakers switch languages inside one recording, use multi-language mode rather than picking a single language. It covers 16 languages today; outside that set, pick the dominant language of the recording.

Realtime covers fewer languages

Live streaming transcription runs on a curated set of 18 languages. Everything else in the table is available for uploaded audio and video files.

Translation runs after transcription

Translation is a separate step on a completed transcript, not a different way of transcribing. One request can name several target languages at once, and each target is billed separately on audio duration × number of languages. See plans and pricing for the current rates.

Accuracy is not uniform

Support means a language is accepted and transcribed, not that every language performs identically. Widely spoken languages with abundant training data are the most accurate; smaller languages are usable but measurably weaker, and audio quality moves the result more than the language choice does. Test your own audio before committing to a language-critical workflow.

Related pages

Building on the API? The Developer API accepts these same codes, and the API documentation covers the transcription and translation endpoints. For a walkthrough of the product itself, see AI transcription and all product features.

Frequently asked questions

How many languages does WhisperAI support?

WhisperAI transcribes 98 languages from audio and video (101 language codes, because English is available in Global, US, British and Australian variants). A completed transcript can then be translated into 87 target languages (91 codes). Live realtime streaming covers a curated set of 18 languages.

What is the difference between a transcription language and a translation language?

A transcription language is one WhisperAI can listen to and turn into text. A translation target is a language an existing transcript can be rendered into afterwards. The two lists overlap heavily but are not identical — a handful of languages can be transcribed but are not offered as translation targets, and Albanian is offered as a translation target only.

Do I have to tell WhisperAI which language the recording is in?

No. Automatic language detection is the default and is right for the overwhelming majority of files. Set the language explicitly when you already know it and the audio is noisy, heavily accented, or short — a fixed language stops detection from drifting.

What if a recording contains more than one language?

Switch it to multi-language mode, which is built for code-switching — speakers moving between languages within the same recording. That mode currently covers 16 languages.

Is accuracy the same in every language?

No. Accuracy is highest in the widely spoken languages with the most training data — English, Spanish, French, German, Portuguese, Italian, Japanese and Chinese among them — and lower in languages with less available speech data. Audio quality, accent, crosstalk and background noise move accuracy more than the choice of language does. If a language matters commercially to you, test it on your own audio first.

Which languages does the Developer API support?

The same lists. Pass a language code on the transcription request (or omit it for auto-detect), and pass an array of target language codes to the translation endpoint. The codes on this page are exactly the values the API accepts.

Can I translate a transcript into several languages at once?

Yes. One translation request can name any number of target languages. Each target language is billed separately, on audio duration multiplied by the number of languages.

99 languages

Transcribe in the language you actually work in.

Upload a file, let WhisperAI detect the language, and translate the transcript afterwards if you need it in more.

WhisperAI
Powered byOpenAI

Professional AI-powered voice transcription and translation platform.

Product

  • Features
  • Plans & Pricing
  • Whisper API
  • Cloud Sync
  • WhisperAI MCP
  • For Enterprise
  • AI Transcription
  • Whisper Transcription
  • Speech to Text
  • Chrome Extension

Resources

  • Blog
  • All Guides
  • Help Center
  • Audio to Text
  • How-to Tutorials
  • For Education
  • For Content Creators
  • For Sales & Marketing
  • For Personal Productivity
  • API Documentation

Compare

  • Compare transcription tools
  • vs Otter.ai
  • vs TurboScribe
  • vs Rev
  • vs Fireflies
  • vs Descript
  • vs Deepgram
  • vs OpenAI Whisper

Popular Guides

  • Podcast Transcription
  • Video Subtitles
  • Legal Transcription
  • Medical Transcription
  • How to Transcribe Audio
  • Transcribe M4A Files

Languages

  • English
  • Spanish
  • French
  • German
  • Portuguese
  • Japanese
  • Chinese
  • Arabic
  • Hindi
  • Russian
  • All supported languages

Company

  • About Us
  • WhisperAI Security
  • Contact Us

Legal

  • Privacy Policy
  • Terms of Service
  • Cookie & Privacy Setting

Follow us on

  • X
  • Instagram
  • LinkedIn

© 2026 WhisperAI Technology Inc. All rights reserved. WhisperAI is a trademark of WhisperAI Technology Inc.