Language guides
Safe normalization
Canonical equivalence, not automatic repair.
POST /v1/text/normalize applies NFC to the complete input and returns input,
output, changed and profile unicode-nfc.
{ "text": "က" }The operation is deterministic and idempotent for the supported repertoire. It
does not apply compatibility normalization, repair spelling or orthographic
order, strip whitespace or controls, or convert Zawgyi. There are no strict or
search profiles in this API.
NFC can change code-point count and ordering. Compute subsequent offsets from the normalized output; no input-to-output offset map is returned. Canonically normalized text is not necessarily valid Burmese.