comfyui-art-venture

Pick a ratio, get width and height

Reach Claude through your AWS account

Mistral through your AWS account

AV_CheckpointMerge — the deprecated model-blend node (use ModelMergeSimple)

Bundle model names into one PIPE

Bake a merged model into a .safetensors

Wire Anthropic's Claude into ComfyUI

One node instead of three

Both conditionings plus step timing

Chain multiple ControlNets on one wire

Multi-ControlNet without the fiddly bits

Load a ControlNet, by name or override

AV_ControlNetPreprocessor — the deprecated all-in-one preprocessor (and what to use instead)

Connect ComfyUI to Google Gemini

The settings node for the art-venture LLM chain

The node that actually runs the LLM call

One prompt in, text out

Build the conversation you send to the model

Stack multiple LoRAs from text

AV_LoraListStacker — build a LoRA stack from a text list, and chain it onto others

A LoRA loader with a mute switch and an override

Image-gen settings for Gemini's image model

Connect GPT models to your ComfyUI graph

One key, every LLM, into your workflow

Unpack a PIPE's model names

Unpack a PIPE back into text, image, and mask

Bundle your prompts into one PIPE wire

Load Segment Anything for precise masking

Load a VAE, with a name-override for automation

BLIPCaption — turn an image into a plain-English caption for img2img and prompting

Load a BLIP checkpoint you already have on disk

One named true/false switch for the rest of your graph

Pick a checkpoint name, as a string

Recolor a grayscale image using another image's color

ColorCorrect — a one-node color grade for images inside ComfyUI

DeepDanbooruCaption — tag an image with Danbooru tags (the old-school interrogator)

Override one piece of a shared generation manifest

DownloadAndLoadBlip — fetch and load the BLIP captioning model in one node

Grab a background-removal model

The Fooocus look, without leaving ComfyUI

The Fooocus sampler, split across passes

Pull a true/false flag out of a JSON blob

Pull one numeric value out of a JSON blob

Pull a number out of JSON by key

Drill into nested JSON

Encode an image once for fast masking

Pull one string out of a JSON blob

Layer one image over another using its alpha

Write a mask into an image's R/G/B/A channel

Pull one R/G/B/A channel as a mask

ImageGaussianBlur — a plain Gaussian blur for images in your ComfyUI graph

Pick one of up to four images by a number

Turn one image into a batch

Shrink images to a target size, with optional crop

ImageScaleDownBy — shrink an image by a simple multiplier

ImageScaleDownToSize — clamp an image's longest (or shortest) side to a pixel size

ImageScaleToMegapixels — resize an image to a target pixel budget, not a fixed width

Load a background-removal model you already have on disk

Cut the background out with ArtVenture's built-in model

LaMaInpaint — erase an object from an image without a diffusion model

Fetch a hosted image and use it as a mask directly

LoadImageFromUrl — pull images straight into ComfyUI from a URL or base64

LoadJsonFromText — parse a block of JSON text into a usable JSON object

Pull live JSON into a workflow

Load LaMa for clean object removal

MergeModels — blend two models by a ratio inside your ComfyUI graph

Remap a value from one range into another

Paste your inpaint crop back onto the original

Composite an inpaint in latent space

PrepareImageAndMaskForInpaint — the A1111-style inpaint prep ComfyUI doesn't have natively

Make a scannable QR image for QR art

A random decimal between two bounds

RandomInt — a random integer generator for seeds and random choices

See what SAM encoded

SDXLAspectRatioSelector — pick an aspect ratio, get SDXL-friendly width and height

SeedSelector — flip between a locked seed and a random one without rewiring

StringToInt — convert a text value into an integer ComfyUI can use

Turn text into an INT and a FLOAT

Cheap wildcard-style prompt randomizing

Pick a string based on a condition