Multilingual Speech-to-Text

Speech-to-Text in every language your users speak

Transcribe and understand speech through a single API — real-time voice agents and batch transcription, with automatic language detection and code-switching built in.

Universal-3.5 Pro

Try uploading a file in any language — we'll detect it automatically.

import assemblyai as aai

aai.settings.api_key = "YOUR_API_KEY"

audio_file = "https://assembly.ai/wildfires.mp3"

config = aai.TranscriptionConfig(
    speech_models=["universal-3-pro", "universal-2"],
    language_detection=True,
    speaker_labels=True,
)

transcript = aai.Transcriber().transcribe(audio_file, config=config)

if transcript.status == aai.TranscriptStatus.error:
    raise RuntimeError(f"Transcription failed: {transcript.error}")

print(transcript.text)

for utterance in transcript.utterances:
    print(f"Speaker {utterance.speaker}: {utterance.text}")

Speech-to-Text for Voice AI in 99 Languages

Our AI speech-to-text transcription supports every language below—just select a language and upload your audio file.

🇿🇦 Afrikaans 🇦🇱 Albanian 🇪🇹 Amharic 🇸🇦 Arabic 🇦🇲 Armenian 🇮🇳 Assamese 🇦🇿 Azerbaijani 🇷🇺 Bashkir 🇪🇸 Basque 🇧🇾 Belarusian 🇧🇩 Bengali 🇧🇦 Bosnian 🇫🇷 Breton 🇧🇬 Bulgarian 🇲🇲 Burmese 🇪🇸 Catalan 🇭🇷 Croatian 🇨🇿 Czech 🇩🇰 Danish 🇳🇱 Dutch 🇬🇧 English 🇪🇪 Estonian 🇫🇴 Faroese 🇫🇮 Finnish 🇫🇷 French 🇪🇸 Galician 🇬🇪 Georgian 🇩🇪 German 🇬🇷 Greek 🇮🇳 Gujarati 🇭🇹 Haitian Creole 🇳🇬 Hausa 🇺🇸 Hawaiian 🇮🇱 Hebrew 🇮🇳 Hindi 🇭🇺 Hungarian 🇮🇸 Icelandic 🇮🇩 Indonesian 🇮🇹 Italian 🇯🇵 Japanese 🇮🇩 Javanese 🇮🇳 Kannada 🇰🇿 Kazakh 🇰🇭 Khmer 🇰🇷 Korean 🇱🇦 Lao 🇻🇦 Latin 🇱🇻 Latvian 🇨🇩 Lingala 🇱🇹 Lithuanian 🇱🇺 Luxembourgish 🇲🇰 Macedonian 🇲🇬 Malagasy 🇲🇾 Malay 🇮🇳 Malayalam 🇲🇹 Maltese 🇨🇳 Mandarin 🇳🇿 Maori 🇮🇳 Marathi 🇲🇳 Mongolian 🇳🇵 Nepali 🇳🇴 Norwegian 🇫🇷 Occitan 🇦🇫 Pashto 🇮🇷 Persian 🇵🇱 Polish 🇵🇹 Portuguese 🇮🇳 Punjabi 🇷🇴 Romanian 🇷🇺 Russian 🇮🇳 Sanskrit 🇷🇸 Serbian 🇿🇼 Shona 🇵🇰 Sindhi 🇱🇰 Sinhala 🇸🇰 Slovak 🇸🇮 Slovenian 🇸🇴 Somali 🇪🇸 Spanish 🇮🇩 Sundanese 🇹🇿 Swahili 🇸🇪 Swedish 🇵🇭 Tagalog 🇹🇯 Tajik 🇮🇳 Tamil 🇷🇺 Tatar 🇮🇳 Telugu 🇹🇭 Thai 🇨🇳 Tibetan 🇹🇷 Turkish 🇹🇲 Turkmen 🇺🇦 Ukrainian 🇵🇰 Urdu 🇺🇿 Uzbek 🇻🇳 Vietnamese 🏴󠁧󠁢󠁷󠁬󠁳󠁿 Welsh 🇮🇱 Yiddish 🇳🇬 Yoruba

Capabilities

Transcription and understanding in every language

Enable any combination of features with a single request. Mix, match, and scale as your application grows.

Code-switching

When speakers switch languages mid-conversation, the model follows seamlessly — preserving exactly what was said instead of forcing everything into one language.

automatic language detection

Identify the spoken language automatically and transcribe without pre-selecting one — ideal when your audio spans regions, dialects, and speakers.

Translation

Transcribe and translate across 99+ languages in the same API request.

Speaker identification

Label speakers using audio context, with support for custom profiles and multi-speaker conversations.

Entity detection

Identify names, organizations, locations, dates, and other entities across every transcript.

PII redaction

Automatically remove sensitive data like credit cards, phone numbers, and personal information from transcripts and audio.

Custom spelling & key terms

Boost recognition of brand names, product terms, and domain-specific vocabulary in any language.

Custom formatting

Normalize dates, phone numbers, and email addresses to machine-readable formats automatically.

Multi-channel transcription

Transcribe separate audio channels independently — ideal for multilingual call center recordings.

Multilingual Speech-to-Text

Built for a multilingual world

From call centers in Mexico City to media companies in Tokyo—AssemblyAI provides one API for transcription, understanding, and scale in every language.

One API, no per-language setup

Transcription, Speech Understanding, and Voice Agents through one integration — cloud or self-hosted, fully compliant. No language-specific infrastructure to build or maintain.

A map of regions and dialects AssemblyAI transcribes

Accuracy across every dialect

Transcribe audio in its original language with industry-leading accuracy — from Canadian French to Brazilian Portuguese to UK English, even with cross-talk and background noise.

Flat, predictable pricing

Pay the same rate for every language and dialect, with transcription, summaries, sentiment, and streaming included. No per-language surcharges.

Common questions

: AssemblyAI's Universal speech models support 99 languages for transcription, with automatic language detection so you don't have to specify the language in advance. Each language has a dedicated page with details on accuracy, capabilities, and supported models.
: Yes. Automatic Language Detection identifies the spoken language and transcribes it without you pre-selecting one — useful when your audio spans multiple regions, dialects, or speakers.
: Code-switching is when a speaker shifts between languages within the same conversation or sentence. AssemblyAI follows these switches seamlessly and transcribes exactly what was said, rather than forcing everything into a single language.
: Yes. You can transcribe audio in its original language and translate across 99+ languages in a single API request, alongside summarization, sentiment analysis, and other Speech Understanding features.
: Yes. Alongside batch transcription, AssemblyAI offers real-time streaming models for multilingual voice agents and live applications, so you can build low-latency experiences in the languages your users speak.
: No. Pricing is flat across every language and dialect — you pay the same rate whether you're transcribing English, Spanish, Japanese, or any other supported language.