script is required only when there is none. Changing the text or voice does not record new audio: call the matching /render endpoint afterward, or upload audio. The response is the stored record; read the speech word for its recording state and playback link.Authorization: Bearer ********************curl --location --request PUT 'https://awqacore01.crexendocloud.com/ns-api/v2/speech-words/hour-10' \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data '{
"script": "ten o clock"
}'{
"filename": "hour-10",
"domain": "*",
"language": "en_us",
"script": "ten o clock",
"voice-engine": "WaveNet",
"voice-id": "en-US-Wavenet-C",
"voice-language": "en-US",
"source-type": "tts",
"content-hash": "3f786850e387550fdab836ed7e6dc881de23001b",
"pre-silence-ms": 500,
"post-silence-ms": 200,
"created-datetime": "2026-09-01T14:30:00+00:00",
"updated-datetime": "2026-10-01T14:30:00+00:00"
}