ComfyUI Workflow
Execute ComfyUI workflows such as image upscaling and video generation.
In Development: This feature is in development. Not ready for use.
Introduction
ComfyUI is an open-source program that allows users to use generative AI such as text-to-image, image-to-image, text-to-video and more through a flexible node-based interface. Heurist SDK abstracts away the complexity of ComfyUI, and provides simple functions in Node.js SDK to execute ComfyUI workflows on distributed GPU miners.
Available Workflows
The SDK currently supports the following ComfyUI workflows:
- Text to Video Generation - Generate videos from text descriptions
- Image Upscaling - Upscale images while maintaining quality
Core Concepts
Each workflow in ComfyUI is represented by a WorkflowTask
. The SDK provides a base class that handles common functionality, which specific workflow implementations extend with their unique requirements.
Data Structures
Each WorkflowTask
represents a task of a specific ComfyUI workflow (identified by a JSON workflow file describing the node structures in ComfyUI). A derived class of WorkflowTask
is defined for each individual workflow, including necessary input parameters.
Base Class
Task Option Type
Task Result Type
Functions
Execute a workflow
Executes a workflow task without waiting for the result.
Returns a Promise that resolves to the task ID.
Query the task result
Queries the result of a previously executed task.
Returns a Promise that resolves to the task result.
Execute a workflow and wait for the result
Executes a workflow and waits until a result is returned. Internally, this function calls queryTaskResult
at a fixed interval, and throws if the task times out.
Cancel a task
Cancels a previously submitted task.
Returns a Promise that resolves to the task ID and message.