Interface STTCapabilities

Describes the capabilities of the STT provider.

Remarks

At present, the framework only supports providers that have a streaming endpoint.

interface STTCapabilities {
    alignedTranscript?: false | "word" | "chunk";
    chatContext?: boolean;
    diarization?: boolean;
    interimResults: boolean;
    keyterms?: boolean;
    streaming: boolean;
}

Properties

alignedTranscript? chatContext? diarization? interimResults keyterms? streaming

Properties

Optional alignedTranscript

alignedTranscript?: false | "word" | "chunk"

Whether this STT supports aligned transcripts with word/chunk timestamps.

  • 'word': Provider returns word-level timestamps
  • 'chunk': Provider returns chunk-level timestamps (e.g., sentence/phrase boundaries)
  • false: Provider does not support aligned transcripts

Optional chatContext

chatContext?: boolean

Whether the STT can natively consume conversation context (see STT._pushConversationItem)

Optional diarization

diarization?: boolean

Whether this STT supports speaker diarization.

interimResults

interimResults: boolean

Optional keyterms

keyterms?: boolean

Whether the STT supports keyterm prompting

streaming

streaming: boolean

Settings

Member Visibility

Theme

On This Page

alignedTranscriptchatContextdiarizationinterimResultskeytermsstreaming