[Topfi]

They've already shared it under a none-commercial use license: https://github.com/tldraw/draw-fast

The TLDraw team, from what I have seen, is really open in sharing their experiments under their own license. Not FOSS strictly, but I feel their licensing approach and decision is fair considering how they fund development, know that it's a contentious topic though for some.

[xienze]

According to that link, you're running the frontend locally but all the work is happening on fal.ai. So the interesting part is not open source.

[steveruizok]

Here’s a resource for more LCM models, which were built over Stable Diffusion. The inference provider is less a design piece than a detail.

https://huggingface.co/latent-consistency