About this project
OmniChain is built as a robust multi-purpose workflow builder and controller for automating tasks with language models used as flexible processors rather than pseudo-humans. It is designed as a hybrid between a classic workflow app (e.g. LangFlow) and a visual programming language.
Short FAQ
Why is the core using NodeJS and not Python?
NodeJS is lightweight and robust in end-user deployment scenarios. Python capabilities are offloaded to OmniChain's external Python module, which is a separate server you can install only if you need it. This keeps the core easy to deploy, while still allowing for Python integration to handle more advanced tasks.
What makes this different from other workflow builders?
Mainly the fact that it leans more towards visual programming than the traditional dataflow style of workflow apps, nodes' ability to update their own contents during runtime to modify how the chain works (this being a persistent exportable modification), and the barebones approach that allows high customizability.
Why the barebones approach?
To make an extendable workflow builder/runner, and still keep it properly maintained over a long period of time without too much risk of updates causing random bugs, it is a good idea to keep the interface and overall logic as clean and straightforward as possible. Thus, OmniChain focuses on being robust rather than flashy.
I don't see any mentions of a team, who's behind this?
OmniChain is a solo project by Filip Filipov, a senior software engineer focused on end-user applications and automation. In other words, all the design, development, documentation, and video tutorials are done by one person. While this is a bit difficult, it allows for a more focused and consistent development process, with no risk of conflicting visions or goals derailing the project.
Can I use this commercially?
OmniChain is open-sourced under the MIT license, so you can freely use it for commercial purposes.