Languages and mid-sentence switching

Choosing your languages

VoiceSnap Pro transcribes 50+ languages. Settings → Languages has two things to set: a primary language, and the shortlist of languages you actually speak.

The primary language is the assumption VoiceSnap Pro starts from when a dictation is short or ambiguous — a two-word search query rarely carries enough signal to identify a language. The shortlist is everything else you want it to recognise.

Why the shortlist matters more than the full list

VoiceSnap Pro auto-detects between the languages on your shortlist, not across all of them. A shortlist of two or three is faster and noticeably more accurate than a shortlist of ten, because the model has fewer ways to be wrong about a mumbled word.

Add a language when you use it, and remove it when a project ends. If you only ever dictate in one language, leave the shortlist at one entry — auto-detection then costs you nothing at all.

Switching mid-sentence

You do not have to announce a switch or press anything. Start speaking German in the middle of an English sentence and VoiceSnap Pro follows, as long as both languages are on your shortlist.

Detection happens at phrase boundaries rather than word by word, so a clause switches cleanly while a single borrowed word usually does not. That is deliberate: the alternative flips language on every loanword and produces a mess.

• “Can you send me the Antragsformular before Friday?” — stays English, keeps the German noun.
• “I'll call them tomorrow. Ich melde mich danach bei dir.” — switches at the sentence boundary and punctuates the second sentence as German.
• Names and product names never trigger a switch. Add them to custom vocabulary if the spelling comes out wrong.

Punctuation follows the language

Automatic punctuation is language-specific, not a set of English rules applied to other alphabets. Spanish questions get their opening ¿, French gets the narrow space before ; : ? and !, German capitalises nouns, and Japanese and Chinese get 。 and 、 with no spaces between words.

Right-to-left languages are inserted in logical order, so Arabic and Hebrew display correctly in any field that supports them. Filler-word removal is language-aware too: it strips the local equivalents rather than only the English ones.

Spoken punctuation and commands in other languages

Command words are localised. Say “nueva línea” in Spanish, “nouveau paragraphe” in French, “neuer Absatz” in German — the English phrases are not recognised while you are speaking another language, which stops an ordinary English-sounding phrase from being eaten as a command.

The full command list per language is in Settings → Languages, next to each entry.

Custom vocabulary across languages

Vocabulary entries apply in every language. That is usually right — names, companies and product names should be spelled the same way whatever sentence they land in.

If a term genuinely differs by language, add both spellings as separate entries with sounds-like values that match how each is pronounced.

What VoiceSnap Pro will not do

It transcribes; it does not translate. Speak French and you get French text. There is no setting that turns speech in one language into writing in another, and adding one would make every mis-detected word a silent mistranslation rather than an obvious typo.

Accuracy also varies by language and accent — the widely spoken languages have far more training data behind them than the rarer ones. If a language matters to your work, test it with the kind of material you actually dictate before you rely on it.