lingua-go

Why we included this project

Knowing what language a piece of text is written in is usually the first step before a Go service can classify content, spell check it, or route it anywhere, and lingua-go handles that step well even when the input is very short, such as a chat message or a single search term. It detects 75 languages using a mix of rule-based and statistical methods, needs no word dictionaries and no connection to an external service, so it runs fully offline inside your binary. That is the real difference from older Go libraries in this space, which tend to lose accuracy on short snippets or when too many languages are in play. Teams routing emails to the right regional desk, triaging multilingual support tickets, or preparing text for spell checking and classification get a compact, dependency-light drop-in rather than a heavy ML stack to operate.