Key Features
Tech Stack
For example, Network (1976) - one of my favourite films - isn't listed on the Wikimedia page, but it's listed on the WikiFlix frontend. I was a little surprised to see that, since AFAIK it's still under copyright. Clicking through, it's trying to embed a copy from the Internet Archive, from which it was taken down because, yes, it's still under copyright.
https://www.wikidata.org/wiki/Help:WikiFlix
Network's WikiData entry is https://www.wikidata.org/wiki/Q572165, which includes the Archive link and lacks a copyright status. This is also linked to from the barcode icon at the top of the movie's WikiFlix page: https://wikiflix.toolforge.org/#/entry/572165
Took about 2 minutes with no prior experience to remove the dead IA link and add the copyright status attribute to the WikiData item.
And then to take a movie like The Matrix, or Star Wars, and then recreate it, shot for shot, as closely as I can, using clips from Public Domain movies.
Especially if I can pick, like, "this actor in this public domain movie, or movies, would be a good Neo, or Luke." And it turns out that he's in enough shots, and poses, that I'm able to remake all of that character's scenes, from Public Domain footage of this other actor.
https://lintangwisesa.github.io/MediaPipe-in-JavaScript/inde...
https://github.com/cosyneco/MediaPipe.NET
https://ai.google.dev/edge/mediapipe/solutions/guide
I'd think you'd start with a tool to automatically cut video into scenes, and kind of go from there...
I'm also reminded of those posters in the mall, where from a distance, it's an actor's face. But when you walk up close, it's actually made of hundreds of stills from their films.
Enjoy.
You start with a digital copy of the movie I'm starting from...
I extract the audio...
And then my script downloads the Public Domain movies...
...and I've got a hard-coded list of cuts to take from each of the Public Domain movies...
...and combine it with the movie audio?
My script is just a few hundred KB.
As long as the movie you start from is pretty well in audio synch with the version I start from...
Heck, I could probably even compensate for that, too... A little bit of analysis to listen for the first words spoken...
Qwen3-VL can scan two-hour videos and pinpoint nearly every detail - https://news.ycombinator.com/item?id=46094606 - December 2025
Not affiliated with Hacker News or Y Combinator. We simply enrich the public API with analytics.