Reverse engineered prompt

GitHub

I want this repo turned into a working app I can actually use to reconstruct a 3D scene from a normal RGB video, ideally in real time when possible. Please get the project installed, make sure the model weights download automatically, and wire up the existing web UI so I can launch both the offline mode and the online mode from the command line and upload either a video, a folder of images, or sample data.

Please make the common path smooth for a beginner. I should be able to run the demo on the provided Replica example, try the outdoor sample, and then swap in my own captured footage with clear instructions on where to put files and what command to run. Save reconstruction results in an easy to find results folder, and make the visualization work too, including the incremental view if that is supported.

If anything is brittle, fix it and add a short README section with the exact steps. Feel free to check current docs online if you need to.

Have a live product UI? Try website reverse