Documentation Index
Fetch the complete documentation index at: https://dev.docs.inworld.ai/llms.txt
Use this file to discover all available pages before exploring further.
Examples
// Voice Agent example: Extracting audio with metadata
class AudioExtractorNode extends CustomNode {
async process(context: ProcessContext, input: DataStreamWithMetadata): Promise<GraphTypes.Audio> {
const metadata = input.getMetadata();
// Check metadata to determine if processing is complete
if (!metadata.interaction_complete) {
throw new Error('Interaction not complete');
}
// Extract completed audio from metadata
return metadata.completed_audio as GraphTypes.Audio;
}
}
Constructors
Methods
Constructors
constructor
new DataStreamWithMetadata(streamOrWrapper: object, metadata?: { [x: string]: any; }): DataStreamWithMetadata
Parameters
The stream to wrap. Can be a raw stream or any typed stream wrapper.
Metadata object containing contextual information. Common fields include iteration, completed, elementType, and any custom application data.
Returns
DataStreamWithMetadata
Methods
getMetadata(): { [x: string]: any; }
Returns
{ [x: string]: any; }
getElementType
Gets the element type of the stream if specified in metadata. Indicates what type of data elements the stream contains.
Returns
string
toStream
toStream(): AudioChunkStream | TextStream | ...
metadata.elementType (most reliable, set when wrapping) 2. stream.type property (fallback)
Returns
AudioChunkStream | TextStream | ...
toTextStream
toTextStream(): TextStream
Returns
TextStream
toAudioChunkStream
toAudioChunkStream(): AudioChunkStream
Returns
AudioChunkStream
toContentStream
toContentStream(): ContentStream
Returns
ContentStream
toTTSOutputStream
toTTSOutputStream(): TTSOutputStream
Returns
TTSOutputStream
