Interface: AudioMetadata
@kortexya/reasoninglayer / AudioMetadata
Interface: AudioMetadata
Section titled “Interface: AudioMetadata”Defined in: src/types/ingestion.ts:2456
Transcription metadata of an audio ingest.
Properties
Section titled “Properties”durationSecs
Section titled “durationSecs”durationSecs:
number
Defined in: src/types/ingestion.ts:2458
Audio duration in seconds (0 when the transcriber did not report it).
language?
Section titled “language?”
optionallanguage?:string|null
Defined in: src/types/ingestion.ts:2460
Detected or observed language.
model:
string
Defined in: src/types/ingestion.ts:2462
Model identifier the transcriber reported.
originalFileId?
Section titled “originalFileId?”
optionaloriginalFileId?:string|null
Defined in: src/types/ingestion.ts:2469
The retained original’s id; absent for a parse-only request. Reuse it as
{ stored: { id } } in a later AudioSource.
sourceFilename?
Section titled “sourceFilename?”
optionalsourceFilename?:string|null
Defined in: src/types/ingestion.ts:2464
Original filename.
speakerMap
Section titled “speakerMap”speakerMap:
Record<string,string>
Defined in: src/types/ingestion.ts:2475
Display name per diarized speaker cluster (cluster id as a string →
name), including the Speaker N fallback. The keys are data, not
schema, so they reach the caller unchanged.
transcribeTimeMs
Section titled “transcribeTimeMs”transcribeTimeMs:
number
Defined in: src/types/ingestion.ts:2477
Transcription wall-clock time in milliseconds.
warnings
Section titled “warnings”warnings:
string[]
Defined in: src/types/ingestion.ts:2479
Non-fatal warnings (CPU fallback, LLM unavailable, partial diarization).