After exploring some tools like Rapidminer, I am wondering about integrating these features within KNIME’s Python nodes:
Using a separate Python virtual environment for each node, managed by a configuration YAML file for installing/removing Python packages.
Securely encrypting and using API keys within the Python node.
Welcome to the KNIME forum. You can use the Conda Environment Propagation to store specific python environments
Thank you for your response. The current Conda Env Propagation requires loading an existing environment with its packages. The aim is to enable uploading/attaching a file with the package list to the node, allowing fresh and quick creation of a new environment from this file.
Hi, @knime2456 
I guess your idea of automatic installation of dependencies with just a configuration file might be easier to implement with uv’s Declaring script dependencies feature.
I believe this is a really, really great idea, I like it. It provides more flexible control than conda env propagation.
Yes, and the speed of uv would make very fast and convenient as well. uv is really a game changer in managing Pÿthon projects compared to using heterogenous tools like conda, venv, pip…
It could also allow to use multiple Python versions in the same workflow without interference or having to manage a hell of dependencies.