videosphere
    Preparing search index...

    Result of peeling already-emitted caption text off a cumulative transcript.

    type CumulativePrefixMatch = {
        remaining: string;
        match: "exact" | "normalized" | "words" | "none";
    }
    Index

    Properties

    Properties

    remaining: string

    Text after the matched prefix (may be empty).

    match: "exact" | "normalized" | "words" | "none"

    How the prefix was matched. none means the provider revised too hard to align — callers must not re-emit the whole transcript as new finals.