Skip to main content
Replacements provide the ability to define words or phrases that should be returned in place of the standard output by the speech-to-text model. Most common use case for this feature is abbreviation handling; however, it can also support formatting styles for keyterms and otherwise provide organizations or users control over how certain words and phrases should be returned.

Feature availability:

/transcribe

/streams

/transcripts

Don’t like how Corti STT outputs certain words, phrases, or acronyms? Define replacements to have terms (single words or multi-word phrases) replaced in final text output with your preferred style.

​
How It Works

Replacements are executed as a final step before returning transcript messages. This means are not present in interimResult transcripts and occur after command detection. Replacement configuration is case insensitive, meaning defined casing is ignored and system will preserved casing of word being replaced. This is different than keyterms, which are case sensitive. Word to be replaced must be a (case insensitive) exact match in order to match replacement rule. This means that “BID” and “B.I.D.” would be two separate entries for replacement with “twice daily”; however, “bid” and B.I.d. would be considered equivalent to “BID” and “B.I.D.”, respectively.
Replacement configuration is limited to 1,000 items per connection.
Replacements are made within the “text” entity of the “transcript” response; therefore, no special client-side handling should be required for receiving or handling of replacement text.

​
Example Use Cases

Use of replacements is optional for dictation configuration. Please contact us to report errors, or for more information on this feature.