Hi,

I just built Chainything, a Rust framework and UI designed to easily generate and execute DAG (Directed Acyclic Graph) workflows using reusable components.

The project is designed to be highly modular and is composed of a few key elements:

• The Core Library: A processor-based library that allows you to define and run a pipeline programmatically (e.g., Load Image -> Apply Resize -> Run Convolutional Network Inference -> Send the result to a database). It handles topological sorting via Kahn's algorithm and dynamic data transfer between nodes.

• The UI (egui): A visual, no-code node editor interface built to create, configure, and run your pipelines graphically.

• AI Agent Integration: You can use a natural language prompt to describe your needs, and the agent will automatically generate and connect the pipeline for you in the UI.

I have implemented some basic components (processors) to get started, and I plan to add more in the coming days/weeks. However, I believe that what will make this tool truly powerful is community contributions.

I would love to get your feedback! Since the architecture is designed to make it easy to create new custom processors, I am very interested in your thoughts on:

• The overall architecture (especially how the Processor trait and data passing via type erasure are designed)

• Code quality and whether the patterns used are idiomatic Rust

• Any features or specific processors you'd like to see

Repository & Demo: GitHub - Bessouat40/chainything: Chain operations on images, text, data, and more · GitHub

Thank you in advance for your time and feedback!

1 post - 1 participant

Read full topic