Here's a sample. The Whisper backend covers 100+ languages, all processed locally.
Widest language coverage. Handles major languages plus many regional ones.
Top 25 languages, optimized for Neural Engine. Fastest option if your language is supported.
English-only model tuned for speed. Best latency for English dictation.
Uses macOS system speech recognition. Requires macOS Tahoe or later.
Pick your language upfront for best accuracy. Auto-detect works, but explicit selection is more reliable.
Whisper figures out what language you're speaking. Handy when you switch languages mid-workflow.
Different language for each app. English in VS Code, French in Pages, German in Mail. Switches automatically.
Speak in one language, get text in another. The on-device LLM handles translation without touching the internet.
With the Whisper backend, Whisperer supports 99+ languages — virtually every major language and many regional ones. NVIDIA supports 25 languages, with an English-only optimized model also available.
Yes. Whisperer's AI Translate mode can translate your dictation to any target language, entirely offline. Speak in Spanish, get English text — or any other language combination.
Yes. With per-app profiles, you can configure a different dictation language for each application. For example, dictate in English in VS Code and French in Pages.
Explicit language selection is recommended for best accuracy. Auto-detection works but may occasionally misidentify the language, especially for short utterances or languages with similar phonetics.