About Language Detector
Language Detector identifies the language of any text, from English and Spanish to Swahili, Tagalog or Kazakh. It also gives you the standard ISO codes, the writing system and the closest alternative matches.
How it works
Every language has a statistical "fingerprint" made of the three-letter sequences, or trigrams, that appear most often in it. English is full of "the", "ing" and "and", while Spanish favors "que", "los" and "ión". The tool breaks your text into trigrams and compares them with profiles of more than 180 languages using the open-source franc library. The closest profile wins.
Before comparing, the detector checks the script (Latin, Cyrillic, Arabic, Devanagari, Han and others). This narrows the candidates immediately, which is why text in Japanese or Greek is recognized almost instantly.
What you can use it for
- Translation: find out which language a message is in before you translate it.
- Customer support: route emails and tickets to the right team.
- Developers: get the correct ISO code for
langattributes, APIs and localization files. - Data cleaning: check the languages in a dataset or survey export.
Tips for the best results
- Give it at least one full sentence. Single words and names are often ambiguous.
- Closely related languages, such as Spanish and Portuguese or Serbian and Croatian, can score close together. Check the alternatives list.
- For mixed-language text, test each part separately. The detector returns the dominant language.
- Remove URLs, code and numbers, which add noise without telling the tool anything about the language.