100+ languages with Whisper
Dictate in over a hundred languages with automatic language detection, or speak any of them and have Whisper translate your speech straight into English text.
Whisper dictation on your Mac runs with no server in the loop. oto runs Whisper models locally through WhisperKit, and also bundles Parakeet v3, so you pick the engine per mode. Choose Whisper for language breadth, or Parakeet for fast, English-focused dictation, and keep every word on the machine either way.
If you searched for Whisper by name, you already know why it's good. oto's job is making it feel like a native Mac feature instead of a science project.
Dictate in over a hundred languages with automatic language detection, or speak any of them and have Whisper translate your speech straight into English text.
Parakeet v3 is bundled and fast, tuned for everyday English dictation. Long-form meeting recording runs on it by default, and you can switch engines in Settings anytime.
Choose smaller Whisper models when you want speed, larger ones when you want maximum accuracy. Switch inside the app whenever the job changes.
Pick a model size in oto's settings: lighter for snappy dictation, heavier for accuracy on tricky audio.
Whisper transcribes on your Mac as you talk, in any language it knows, and it knows a lot of them.
The transcription pastes at your cursor and lives in your searchable history, exportable as plain text, Markdown, or JSON.
Whisper and Parakeet are two of up to three local engines in oto, alongside Apple's on-device speech engine, and you can switch freely. Two honest details: the 100+ language count is specific to Whisper, and translation goes one way, speech into English text, not between arbitrary languages.
Whisper if you need language breadth or speech-to-English translation. Parakeet if you're dictating in English and want speed, it's the engine behind long-form meeting recording. You can switch anytime in Settings, and both run on-device.
Yes, into English: speak in any language Whisper supports and get English text back. It's speech-to-English translation, not a general-purpose translator.
One $8 purchase gets you Whisper on your own hardware, plus two more local engines when you want them.