Is this possible to run locally and just have aliases in the CLI? I have a bunch of VHS -> my machine captures of now dead martial arts instructors' seminars -

Is this possible to run locally and just have aliases in the CLI? I have a bunch of VHS -> my machine captures of now dead martial arts instructors' seminars - and I have a decades long desire to digitize then mocap them - and I found some tools for building AI rigs for games/deepfakes - but I would love to be able to Command the Vids into the formats I need to ingest them to such rigging tools - which do a poor job... (Anyone Recall Poser3D? That was a first attempt... Manual rigging...) IDEALLY: We need an ai | (we need a PIPE TO NEXT AI/LLM operator); ... I want to tie all these and more into a string: https://i.imgur.com/XIVz05g.png https://i.imgur.com/XIVz05g.png ">>> *take this vid from 3:33 to 4:14, make a PNG for the first frame, normalize audio, stabilize video and create individula frames cropped to just the humans at 1:1 4k, build an index and create a log file detailing your efforts then | {next CLI LLM function to create AI rig, ALA Scenario.com}" --- EDIT: uh, I uploaded a file [0] then asked it to rotate 90 [1]... this process was really slow to do a transform (upload, select operation, wait, image downloaded... just seemed extremely wasteful in all aspects of energy - time, energy, money, compute. :-( Not denigrating the project - just that its much better to do simplt transforms local and just know how to do that (it makes the world a better place)... But an AI piping to throw vids at and have it effective normalize and prep vids for ingestion is great - but again, want to do that local - until its a library of services that you want from a cloud, such as a use case - where, if you were to use GCP's Visiual stream labeling tooling could save you a bunch of efficiency on speed of AI processing - if you can funnel a stream through this first - then have that piped to GCP vision, you may end up with cleaner, smaller frames resulting in higher accuracy. --- @YenRabbit: >"..quite frequently.." May you please publish some of these and a use case that drove you to GPT`ing them?