Question 1

How does language detection work?

Accepted Answer

The tool uses the franc library which analyzes trigram frequency profiles. Every language has a characteristic distribution of 3-character sequences. By comparing your text's trigram frequencies against known language profiles, it ranks the most likely languages by confidence.

Question 2

How many languages can it detect?

Accepted Answer

The franc library supports detection of over 400 languages including all major European, Asian, Middle Eastern, and African languages. Accuracy is best for texts of 100+ characters.

Question 3

Why is confidence low for short texts?

Accepted Answer

Short texts have fewer trigrams, making statistical comparison less reliable. For best results, paste at least 50–100 characters. Single words or very short phrases may show low confidence across all candidates.

Question 4

Is my text sent to a server?

Accepted Answer

No. Language detection runs entirely in your browser using the franc JavaScript library. Your text never leaves your device.

Question 5

Can it distinguish between similar languages?

Accepted Answer

Yes, but accuracy varies. Closely related languages like Serbian and Croatian, or Spanish and Portuguese, can be confused on short texts. Longer text samples improve accuracy significantly.

Spracherkennung

Häufig gestellte Fragen