v3

latestOpenAPI 3.0.02026-07-318685204.9 KB

Enhancements

Update a speech dictionary by ID

The terms list is replaced as a whole. Terms are cleaned up automatically (whitespace normalized, empty and duplicate entries dropped); terms that cannot be fixed automatically (forbidden characters ";" or "/", more than 6 words) are rejected with invalid_value.

patch/enhancements/speech/dictionaries/{id}

Path parameters

idstring required

Speech dictionary ID

Request body

namestring

Name of the speech dictionary

languagestring

IETF BCP 47 language tag (e.g. "en", "en-US")

termsstring[]

List of terms (names, proper nouns, acronyms, specialized terminology) the speech recognition checks against while captioning the source audio. One term per entry, at most 6 words per term; a term must exactly match what is said in the stream. Optionally append "sounds like" phonetic spellings to a term with a pipe, separating multiple pronunciations with a comma: "D&I|dee-and-eye,dee-n-eye". Entries are cleaned up automatically: whitespace is normalized, spaces around "|" and "," are removed, and empty or duplicate entries are dropped. Entries containing ";" or "/" (not accepted by the speech recognition) or more than 6 words are rejected.

Example request

{
  "terms": [
    "THEOlive",
    "A. Väyrynen",
    "D&I|dee-and-eye,dee-n-eye"
  ]
}

Response

Updated speech dictionary

Example response

{
  "data": {
    "terms": [
      "THEOlive",
      "A. Väyrynen",
      "D&I|dee-and-eye,dee-n-eye"
    ]
  }
}