Introduction

What are Custom Nodes?

Custom nodes are the bridge between your ideas and Blobit’s visual environment. They allow you to extend the capabilities of the application by writing your own Python logic for processing and defining your own LiteGraph.js interfaces for interaction.

This hybrid architecture ensures you get the best of both worlds: a responsive, visual interface in the browser and the raw power of Python’s ecosystem for heavy computation, AI, and data manipulation.

Note

Blobit nodes use a specialized BaseNode abstraction class. This standardizes how data frames are handled, errors are reported, and inputs/outputs are accessed, ensuring stability across the system.

How does it work?

The architecture is split into two distinct parts:

• Frontend (LiteGraph.js): Runs in your browser. It handles the visual representation of nodes, inputs, outputs, UI widgets (buttons, sliders), and user interaction.
• Backend (Python): Runs on the server. It executes the actual logic. When a node is triggered in the graph, the execution engine calls your node’s execute method with the necessary data.

Key Features

• Frame-Based: All data flows as sequences of frames (even single images), allowing for complex animation logic.
• Hot Reloading: Changes to .js frontend files often reload instantly for rapid UI iteration.
• Dependency Management: Nodes can define their own requirements.txt which Blobit installs automatically.