Skip to main content

Glossary

Definitions of terms used throughout DUTO.

A​

Artifact​

Generated content (images, videos, audio) produced by a workflow execution.

Aspect Ratio​

The proportional relationship between width and height of an image or video (e.g., 16:9, 1:1).

B​

Batch Processing​

Processing multiple items through the same workflow segment using BatchProcessor and BatchEnd nodes.

Batch Scope​

The set of nodes between a BatchProcessor and BatchEnd node. Edges within batch scope are displayed as dotted lines to visually distinguish them.

Brain Node​

An intelligence node that uses LLMs for reasoning, analysis, and content generation. Uses the Smart Brain architecture with runtime edge detection.

Budget Mode​

A cost-saving mode that uses more economical AI models for generation.

C​

Canvas​

The main workspace in the Flow Editor where nodes are placed and connected.

Credits​

DUTO's currency for AI operations. Different operations cost different amounts of credits.

Connection (Edge)​

A link between two nodes that defines data flow in a workflow.

D​

DAG (Directed Acyclic Graph)​

The data structure representing workflows - nodes connected by directional edges without cycles.

Duration​

The length of a video in seconds.

E​

Edge​

A connection between nodes in a workflow, also called a "connection."

Execution​

Running a workflow to produce outputs from inputs.

F​

Flow Editor​

DUTO's visual workflow builder where you create and connect nodes.

Flow Plan​

The generated structure that AI creates when building workflows.

G​

Generation​

The process of creating content using AI models.

H​

Handle​

The connection point on a node where edges attach. Input handles are on the left, output handles on the right. Handles are color-coded by data type: blue (image), purple (video), green (audio), yellow (text), orange (json), gray (any).

Handle Type​

The data type a handle accepts or produces. Connection validation ensures compatible types are connected.

Hero Frame​

A preview image generated for each shot in a storyboard before video generation.

I​

Image to Video​

Converting a static image into animated video using AI.

Inpainting​

AI technique to fill in or replace specific regions of an image.

L​

Library​

The Content Library where media assets are stored and organized.

LLM (Large Language Model)​

AI models like Gemini and DeepSeek used for text reasoning and generation.

M​

Model​

An AI system trained to perform specific tasks (image generation, video generation, etc.).

Movie Editor​

DUTO's video composition tool using FFmpeg.wasm for editing and combining video clips in the browser.

Music Video Mode​

A storyboard mode that synchronizes shots with music beats.

N​

Node​

A building block in the Flow Editor that performs a specific function. DUTO has 50 node types across 9 categories.

Node Run​

The execution state of a node during workflow running.

O​

Orchestrator-Worker Pattern​

Architecture used in Storyboard where a central orchestrator coordinates specialized workers.

P​

Pipeline​

A sequence of processing steps, like the 11-worker storyboard pipeline.

Prompt​

Text instructions given to AI models to guide content generation.

Provider​

A service that hosts AI models (WaveSpeed for media generation, OpenRouter for LLMs, Tavily for research).

R​

Reference Image​

An image uploaded to guide AI generation, especially for character consistency.

Regenerate​

Creating a new version of generated content with the same or modified parameters.

Retry Gate​

A flow control node that allows retrying a section of the workflow if upstream nodes fail, with configurable retry limits.

S​

Shot​

A single segment of video in a storyboard, corresponding to one continuous camera take.

Storyboard​

A visual plan for video content, consisting of multiple shots with descriptions and hero frames.

Style Transfer​

Applying the visual style of one image to another image or video.

Smart Brain​

The architecture pattern used by the Brain node where it inspects connected outputs at runtime, makes a single LLM call, then performs targeted extractions for each output type.

Sync Gate​

A node that waits for multiple inputs to be ready before proceeding.

T​

Text to Image​

Generating images from text descriptions using AI.

Text to Video​

Generating videos from text descriptions using AI.

Top-Up​

Additional credits purchased beyond subscription allocation.

U​

Upscaling​

Increasing the resolution of an image or video while maintaining or improving quality.

V​

Veo​

Google's AI video generation model family, including Veo 3.

Video Foley​

Generating sound effects for video based on visual content.

W​

Workflow​

A complete graph of connected nodes that accomplishes a task in the Flow Editor.

Worker​

A specialized AI process in the storyboard pipeline (e.g., Cinematographer, Lighting).

Abbreviations​

AbbreviationMeaning
AIArtificial Intelligence
APIApplication Programming Interface
BPMBeats Per Minute
CPUCentral Processing Unit
DAGDirected Acyclic Graph
FPSFrames Per Second
GPUGraphics Processing Unit
JSONJavaScript Object Notation
LLMLarge Language Model
MP3MPEG Audio Layer 3
MP4MPEG-4 Part 14
PNGPortable Network Graphics
SFXSound Effects
TTSText-to-Speech
UIUser Interface
UXUser Experience
WebPWeb Picture format

Node Categories Reference​

CategoryCountExample Nodes
Flow Control6Start, Sync Gate, Conditional Switch, Retry Gate, Batch Processor, Batch End
Text3Text Prompt, Prompt Expander, Text Preview
Image5Text to Image, Image Upscale, Image Background, Angles
Video15Text to Video, Image to Video, Video Extend, Lip Sync
Audio4Text to Audio, MMAudio, Kling Video to Audio, Video Foley
Intelligence4Brain, Analyzer, Story Generator, Narrative Explorer
Media I/O6Upload Media, API Endpoint, Library, Remote Media, Preview Media
Localization1Variant Generator
Other6Brand Kit, Product Concept, Cinema Studio, Post-It, Preset, Data Feed