Enum Modality

Modality defines the data form a model accepts as input and returns as output. Even for similar generation tasks, models can differ in whether they use text, image, or audio data. The SDK uses this value to validate compatibility in model selection, request mapping, and UI filtering.

Namespace: Glitch9.AIDevKit

Assembly: Glitch9.AIDevKit.dll

Syntax

public enum Modality

Fields

Name	Description
Audio	Represents audio input or output. Includes speech files, waveform data, and synthesized voice output. Used in STT/TTS and realtime voice agent scenarios.
Document	Represents document input. Includes files such as PDF, DOCX, and report-style documents. Commonly used in OCR, document analysis, and RAG preprocessing workflows.
Embedding	Represents embedding-vector data. Typically refers to high-dimensional numeric representations of text. Used for semantic search, similarity scoring, and vector database storage.
File	Represents a generic file input not bound to a specific media type. Useful for uploaded file references and file-based tool workflows. Use with more specific modalities (such as Document or Image) when needed.
Image	Represents image input or output. Covers raster image data such as PNG and JPG assets. Used for vision tasks including image analysis and image generation.
None	Represents an unspecified modality. Used as a default or fallback when the modality is unknown. This is not intended to represent a concrete request/response payload type.
PixelArt	Represents pixel-art specific image input or output. Unlike general images, this assumes pixel-style and resolution-constrained workflows. Used for PixelArts model filtering and feature gating in the SDK.
SkeletonRig2D	Represents 2D character skeleton rig data. Includes structured animation-control data such as bones and joint relationships. Used in pixel-character animation generation and skeleton estimation pipelines.
Text	Represents plain text input or output. Includes prompts, chat messages, and textual generation results. This is the most common modality in standard LLM workflows.
Video	Represents video input or output. Includes moving-image assets and generated video results. Used to determine compatibility for video generation and analysis features.