Updated: Jul 12, 2025
toolVerified: 5 days ago
Other
The FLUX.1 [dev] Model is licensed by Black Forest Labs. Inc. under the FLUX.1 [dev] Non-Commercial License. Copyright Black Forest Labs. Inc.
IN NO EVENT SHALL BLACK FOREST LABS, INC. BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH USE OF THIS MODEL.
About the workflow:
Nodes you will probably want to adjust and use normally.
Init
Load the pictures to be used with Kontext.
Loader
Select the diffusion model to be used, as well as load CLIP, VAE and select latent size for the generation.
Prompt
Pretty straight forward: your prompt goes here.
Switches
Basically the "configure" group. You can enable / disable model sampling, LoRAs, detailers, upscaling, automatic prompt tagging, clip vision UNClip conditioning and IPAdapter. I'm not sure how well those last two work, but you can play around with them.
Model settings
Model sampling and loading LoRAs.
Sampler settings
Adjust noise seed, sampler, scheduler and steps here.
1st pass
The generation process itself with no upscaling.
Upscale
The upscaled generation. By default it makes a factor of 2 upscale, with 2x2 tiled upscaling.
Mess with these nodes if you like experimenting, testing things.
Conditioning
Worthy to mention that FluxGuidance node is located here.
Detail sigma
Detailer nodes, I can't easily explain what does what, but if you're interested, look the nodes' documentation up. I set them at a value that normally generates the best results for me.
Clip vision and IPAdapter
Worthy to mention that I have yet to test how well ClipVision works and IPAdapter's strength when it comes to Flux Kontext.