public/api/types/results library

Classes

AppliedAdapter
One LoRA adapter currently applied to the loaded model.
Audio
Synthesized audio buffer.
AudioChunk
One chunk of a streamed synthesis.
ClassInfo
One class present in a segmentation mask.
DiarizationResult
Who spoke when.
Embedding
One embedding vector and its position in the input list.
GenerationResult
Text, tool calls, and metrics from one generation.
ImageData
One generated image.
ImageResult
Images produced by one generation, plus the sampling parameters used.
LoadedModel
Resident-model ownership handle returned by models.load.
LoraState
Adapters currently applied to the loaded model.
Match
One retrieved chunk and its similarity to the query.
ModelsState
What is loaded and how much storage the SDK is using.
RagCapabilities
RAG-session support of the currently registered backends.
RagResult
Grounded answer plus the chunks it was drawn from.
RagStats
Size of a RAG session's index.
RankedResult
One reranked document pointer.
SDKCapabilities
Installed, packaged, and executable surface of this SDK build.
Segment
A span of audio, in milliseconds from the start of the buffer.
SegmentationResult
Per-pixel class assignment for one image.
SpeakerSegment
One diarized span attributed to a speaker.
SpeechHandle
Handle to one in-flight or completed tts.speak/VoiceSession.say utterance.
StreamingCapabilities
Per-modality streaming support, keyed by namespace name.
StructuredResult
Parsed value plus the raw text and metrics of a structured generation.
SttState
Readiness of the speech-to-text component.
ToolCapabilities
Tool-calling support of the currently registered backends.
Transcription
Transcribed text with timings and language.
UnavailableCapability
Whether a modality/backend/feature is honestly available right now.
VadResult
Voice-activity verdict for one buffer.
Voice
One synthesis voice offered by the loaded model.
Word
One word of a transcription with its timing.

Enums

FinishReason
Why a generation stopped.