You can configure custom sensitive words in the built-in Speech-to-Text (STT) node.
Feature introduction
This feature lets you define a list of sensitive words. Then, the STT node detects these words in real time. If a sensitive word is detected, the system automatically redacts it in the client-side subtitles by replacing it with an asterisk (*).
After a sensitive word is detected, the system passes the redacted content to the large language model (LLM) node.
Configure sensitive words
Prepare a sensitive word file: Add the words you want to block to a TXT file.
NoteThe language of the sensitive words must match the language model that is configured in the console.
Each file can contain up to 500 words. Each word can be up to 10 characters long. The total file size cannot exceed 100 KB.
The file must use
UTF-8encoding. The words cannot contain punctuation or special characters.
Configure the STT node: Upload the sensitive word file to the STT node in the workflow.
Apply the changes: After you upload the file, save the workflow to apply the new configuration.