Boyonodes
Essential nodes for semantic image editing workflows with Kontext, Qwen Image Edit, and HiDream E1.1 models. Features automated iterative editing with Boyo Image Grab, organised dataset creation with paired image/prompt savers, and comprehensive file management tools. Perfect for batch processing, training data preparation, and streamlined semantic editing pipelines. Added Uncensored Qwen VL abliterated model node and downloader!
Nodes (74)
Hand Z-Image a reference photo and it'll match the style — if you bring the right weights
Colour-grade your generations in the graph — 80+ LUTs shipped in the box
The same auto-installer, with the training wheels off — explicit paths, same caveats
A workflow that installs its own nodes, models, and pip packages — handy and a little scary
How long is that audio, exactly? The node that answers before you build around it
Size your video to your audio before the audio is even in the graph
Pad your TTS audio to the video's exact length
The node that re-runs your whole workflow from inside itself — powerful, and a little insane
Five seconds of a voice, one prompt, and it clones it — the TTS generate node
Load the real Chatterbox Turbo TTS into ComfyUI — and yes, the watermark is stripped
A tiny boolean for loop decisions — 'is it done yet?' as a node
See What Sampler Object Is Actually Flowing Through Your Graph
Read the Actual Sigma Schedule Your Scheduler Produced
Force ComfyUI to run things in the order you mean — a wire that's also a wall
It Doesn't Throw Away Your Sampler's Intermediate Guesses Anymore
Draw Your Own Sigma Schedule for Flow Models, and Look at It
One Sampler That Changes Which Solver Runs Each Step
Where loop values come home
The loop start that replicates EasyUse without making you install EasyUse
The one-line math node that keeps infinite video loops in sync
Pull One Prompt Out of a List, Driven by a Counter
Tile-crop a huge image into training patches without leaving ComfyUI
Watch a folder and auto-load the newest image — the loop node that drives iterative editing
Save source-and-edit image pairs for an in-context dataset in one click
The loader half of Z-Image's IP-Adapter — read this before you hunt for weights
An Empty Latent Image with aspect-ratio presets and a zeros mode — that's the whole pitch
Sneak a sampler's output back into the next loop pass without a graph cycle
The latent clone node for when loops start sharing state they shouldn't
The clean way to feed last iteration's latent back into a looped sampler
Step through an input folder one image at a time — dropdown included
Load a whole folder of images as one batch, resized to fit
Load videos from a folder as frames, audio, and VHS video info
Turn every iteration of a bastard loop into one contact sheet
A no-fuss counter for loop logic that doesn't need the heavy loop nodes
Persist every iteration of a loop run to disk, grid or individual
Punch a loop counter back to zero when a workflow section finishes
Check a LoRA config file before it blows up your load step
The logic half of the LoRA pipeline — prompts, strategies, and paths as data
A LoRA dropdown that outputs a filename string, nothing else
Write a LoRA config file without ever opening a text editor
Load three LoRA configs at once — high-noise, low-noise, and the prompts that go with them
The stability bridge between LoRA config data and standard ComfyUI loaders
The Tiny Node That Unblocks DiffuEraser's Video Mask Bug
Integer Math for When Your Loop Math Stops Being Mental
A Plain Multiline Text Box That Does Exactly One Thing
Why Your First Video Chunk Has No Overlap (and This Node Fixes It)
PainterI2V motion with SVI's context preservation
Building a ControlNet Dataset One (Source, Control) Pair at a Time
Save each image next to its prompt as a numbered pair — your dataset starts here
PiD Aspect Ratios Without the Non-Square Seam Headache
The Node That Hands Your Loop Its Next Prompt
Batched Prompts From a Text File, Sequenced for Loops
Different Prompt Every Part of the Video
Drag your prompts onto a timeline and stop fighting one giant video prompt
A Different LoRA for Each Part of Your Video
Width In, Matching Height Out — Always Divisible by 8
Save Before/After Image Pairs With a Single Prefix
Turn a Storyboard's JSON Into 12 Ready-to-Wire Prompt Outputs
12 Storyboard Prompt Outputs, Plus Speech Tracks for Audio Mode
The Node That Writes the Prompt That Writes Your Storyboard
The tiled VAE decoder with a built-in NSFW filter — read the fine print before you use it
VACE Control Without a Single Conditioning Node
The Debug Node That Peeks Inside Your Patched Model
A VAE Decode That Secretly Censors Your Output
Slice Any Video to an Exact Frame Count at an Exact FPS
Stripping the Overlap Frames Out of Chunked Video
How Many Frames Is 30 Seconds of Chunked Video? This Node Knows
Save your video and the prompt that made it, as one unit
The no-fuss ffmpeg video saver for your frame batches
Voice cloning and cleanup without the cloud, via Seed-VC
Stamp your brand on every frame, right before the save
The other half of the loop — where the 'while' actually happens
Open a while loop the EasyUse way, without installing EasyUse
A Mandelbrot zoom for every render, with zero models to download
Essential nodes for semantic image editing workflows with Kontext, Qwen Image Edit, and HiDream E1.1 models. Features automated iterative editing with Boyo Image Grab, organised dataset creation with paired image/prompt savers, and comprehensive file management tools. Perfect for batch processing, training data preparation, and streamlined semantic editing pipelines. Added Uncensored Qwen VL abliterated model node and downloader!