It looked like it worked for me in Forge, once I added the vae and clip etc but the image was blank. I'll try with Comfy later.

https://github.com/lllyasviel/stable-diffusion-webui-forge/discussions/1050

This shows what settings to use at the top of Forge.

With Forge you can just drop Flux checkpoints into the same folder with other checkpoints.

Basically, normal models use keywords and phrases, separated by commas. But with Flux, you basically just fully describe what you want to see, in full detail, as if you were making a request from an actual person. See her prompt on the image of the man holding the sight for reference.