Hi! I've been working on an open-source project called Scrollcase. I started this cause moving Python or ML projects between machines is often painful. You need the right Python version, libraries, native builds, model files, and so on. So instead of making someone rebuild a Python environment on their machine, Scrollcase packages everything once, including Python, locked dependencies, code, and model files.
Then they can verify it, unpack it, and run it. No Python install, no pip install, no Docker, and no dependency to mantain. You just define what goes inside, lock the dependencies, and build a portable, self-contained, verifiable package that's ready to run.
Would love feedback from people who had to package Python, ML, or scientific environments for machines they don't control.
Comments URL: [https://news.ycombinator.com/item?id=49362094](https://news.ycombinator.com/item?id=49362094)
Points: 1