WorkflowListedExternal

Flux 2 Dev - Basic Workflow (Text-to-Image + Reference Support) — v1.0

Here is a starter workflow for the newly released Flux 2 Dev by the ComfyUI team.

TaskMultiCapabilitiesImage to image · Text to imageFamilyFLUXVersionRelease 2439709 — Flux2_Dev_Basic_Workflow.jsonEstimated VRAMEstimate unavailableBgrade

Runtime profile

Access: External workflow
Status: Listed
Version: Flux2_Dev_Basic_Workflow.json
Source: Civitai
Updated: Jul 19, 2026

Source description

Here is a starter workflow for the newly released Flux 2 Dev by the ComfyUI team.

This workflow is designed to get you up and running quickly with the new architecture. It handles standard Text-to-Image generation but also includes a setup for Reference Image Conditioning (Multimodal input).

Workflow Features:

Standard Generation: calibrated for Flux 2 Dev.

Reference Inputs: I have included ReferenceLatent nodes connected to LoadImage inputs. This allows you to prompt using images (similar to IP-Adapter/Variations) combined with text.

Note: As written in the workflow notes, simply Bypass (Ctrl+B) the Reference nodes if you only want to do pure Text-to-Image.

AI-generated commentary

AI-generated explanation based on source and configuration details. Suggestions are clearly labeled.

Purpose

This Flux 2 Dev workflow is for text-to-image generation: it can generate images from a text prompt and optional reference images loaded through LoadImage and ReferenceLatent, with image output handled by SaveImage.

Flow

For text-only generation, bypass the Reference nodes with Ctrl+B.
Suggestion · not verified
For reference-image conditioning, upload images to the LoadImage nodes; the workflow connects them to ReferenceLatent nodes.
Suggestion · not verified

Inputs and outputs

The reference-image path uses LoadImage inputs connected to ReferenceLatent nodes.
The workflow includes CLIPTextEncode, Flux2Scheduler, SamplerCustomAdvanced, and SaveImage nodes.

Setup

Place flux2_dev_fp8mixed.safetensors, mistral_3_small_flux2_fp8.safetensors, and flux2-vae.safetensors, or equivalents, in your ComfyUI models folder.
Suggestion · not verified
Load the workflow, select checkpoints in the Loaders, and queue the prompt.
Suggestion · not verified

Dependencies

The named model files are flux2_dev_fp8mixed.safetensors, mistral_3_small_flux2_fp8.safetensors, and flux2-vae.safetensors.
Listed node types include BasicGuider, CLIPLoader, CLIPTextEncode, EmptyFlux2LatentImage, Flux2Scheduler, FluxGuidance, ImageScaleToTotalPixels, KSamplerSelect, and LoadImage.
Other listed node types are RandomNoise, ReferenceLatent, SamplerCustomAdvanced, SaveImage, UNETLoader, VAEDecode, VAEEncode, and VAELoader.

Configuration

The default resolution is 1024x1024.
The listed base model is Flux.2 D.

Cautions

Before use, check that your ComfyUI installation has the listed node types.
Suggestion · not verified
A license is not specified for this workflow.

Does this need to be edited?

Sources

1 source

CivitaiSourceVersion 2439709

civitai.com/models/2166497

Verified Aug 4

Source excerpts

2 excerpts

NoteAug 4civitai.com/models/2166497

Source context: 1104 downloads · Type Workflows · Base model Flux.2 D

NoteJul 19civitai.com/models/2166497