Text analysis with explicit counting semantics
The result distinguishes Unicode graphemes, code points, word/sentence boundaries, line endings or normalization where those details change the meaning of the count. Browser-local analysis stays live without treating approximate language heuristics as exact facts.