tutorials/flux/flux-2-dev

[giscus[bot]]

tutorials/flux/flux-2-dev

This guide provides a brief introduction to the Flux.2 model and guides you through using the Flux.2 Dev model for text-to-image generation in ComfyUI.

https://docs.comfy.org/tutorials/flux/flux-2-dev

[kalle07]

with at least one image as input:
TextEncodeQwenImageEdit - node works !!!
but not TextEncodeQwenImageEdit_plus !
to attach more image use "Image Stitch" - node
its a bit slower but much more comfortable.

[bebelllm]

the old workflow was easier with several image, why did you get rid of the several input image workflow ?
where to find the old workflow ?

[Xayah-Hina]

Could you please restore the old workflow which enables multiple reference images?

[bebelllm]

I agree. The developpement went backward, more complicated to use, less friendly. The opposite as they promoting. Bad.
So I'm using the old template, the new I can't use it with several images.

[djubre7]

@bebelllm can you link here the old workflow? Thanks in advance!

[bebelllm]

@bebelllm can you link here the old workflow? Thanks in advance!

I think i went on the github on the older workflow files and download it there. Try to search a bit in this direction.

[comfyui-wiki]

Hi @Xayah-Hina,

Flux.2 Dev actually has two workflows available:

1. Single image workflow (current default)
2. Multiple reference images workflow (the one you're looking for)

How to access the multi-image workflow:

Option 1 - Cloud ComfyUI:
Open this URL directly:

http://cloud.comfy.org/?template=image_flux2_fp8

Option 2 - Local ComfyUI:
Add /?template=image_flux2_fp8 to your local ComfyUI URL:

http://127.0.0.1:8188/?template=image_flux2_fp8

(Replace with your actual port if different)

Usage tip:

The workflow uses subgraphs/groups for organization. Simply unpack the subgraph to access and modify the multiple image inputs as needed.

Hope this helps! Let me know if you have any questions.

---

This response was AI-assisted