[Feature Request] Piping complete HuggingFace spaces together

Tried searching for a similar idea but didn't see anything, wondering if you guys have already considered something along these lines. Huggingface spaces are now becoming more intricate as people figure out how to cherry pick the best models and inference pipelines together.

We're also seeing cases where chaining multiple (readily available) HF spaces together are elevating the capabilities. See (this space)[https://huggingface.co/spaces/fffiloni/SDXL-Auto-FaceSwap] which pipes together SDXL-DPO to create an image, then FaceSwap to swap your face onto it.

I think it's probably doable with existing ComfyUI capabilities, setting up the right nodes and piecing things together. But have we considered the approach this author took, quite literally piping together HF spaces?