Find your language.
Choose a transcription model that supports the language you speak. Dictation and Notetaker audio transcription use the model-specific lists below.
Audio transcription
Choose your spoken language and a compatible model in Orca One. Download the model before using it offline. Language support does not guarantee the same accuracy for every voice or recording.
Finnish is supported by Parakeet Multilingual, Parakeet Ultra, Qwen3 ASR, and Qwen3 ASR Large. Parakeet English and Cohere Transcribe do not support Finnish.
Parakeet English / Parakeet Small English
1 language
- English
Parakeet Multilingual / Parakeet Ultra
25 languages
- Bulgarian
- Croatian
- Czech
- Danish
- Dutch
- English
- Estonian
- Finnish
- French
- German
- Greek
- Hungarian
- Italian
- Latvian
- Lithuanian
- Maltese
- Polish
- Portuguese
- Romanian
- Russian
- Slovak
- Slovenian
- Spanish
- Swedish
- Ukrainian
Parakeet Japanese
1 language
- Japanese
Qwen3 ASR / Qwen3 ASR Large
30 languages
- Arabic
- Cantonese
- Chinese
- Czech
- Danish
- Dutch
- English
- Filipino
- Finnish
- French
- German
- Greek
- Hindi
- Hungarian
- Indonesian
- Italian
- Japanese
- Korean
- Macedonian
- Malay
- Persian
- Polish
- Portuguese
- Romanian
- Russian
- Spanish
- Swedish
- Thai
- Turkish
- Vietnamese
Cohere Transcribe
13 languages
- Arabic
- Chinese
- Dutch
- English
- French
- German
- Greek
- Italian
- Korean
- Polish
- Portuguese
- Spanish
- Vietnamese
Cohere Transcribe requires a spoken-language selection and supports the 13 languages listed here. Japanese is not supported by Cohere Transcribe in Orca One. Qwen3 ASR Large uses more memory than Qwen3 ASR; the app shows model requirements before installation.
Beyond transcription
Writing styles
The local S1-mini writing-style model supports English only. For Finnish and other languages, Orca One keeps the transcription without applying a writing style.
Meeting notes and summaries
Meeting audio uses the transcription model and language you select. Summarize sends a request to your chosen connected AI app. Summary languages depend on that app and its model; the transcription lists are not a summary-language guarantee.
Voice launcher
A spoken query is transcribed with a compatible speech model. Matching apps, files, quicklinks, and scripts is a separate step. These lists do not guarantee command understanding in every language; test the names and commands you use.
App interface
These lists describe spoken-language transcription, not translated menus or interface localization.