Quality of Life Nodes for ComfyUI
These nodes aid with batching image processing and maintaining input file names in output files and other quality of life nodes.
Nodes (86)
Add a line without overwriting the rest
Stitch two clips together with ffmpeg
The NOT gate stock ComfyUI doesn't ship
Turn true/false into text ComfyUI's string nodes can use
Load a checkpoint from a raw path, not a dropdown
Bundle toggles for a ComfyDeploy run
Read back everything you've queued
Remember every run_id you fired off
Reset the cache between batches
Wait for a run and pull down its output
Pack decimal values into a ComfyDeploy request
Send images to a ComfyDeploy deployment
Pack whole numbers into a ComfyDeploy request
Five parameter types, one node
The cleaned-up mixed builder
Fire off a ComfyDeploy run from inside ComfyUI
Read back a queued run later
Build the payload for a ComfyDeploy run
Turn a JSON list into a loadable image batch
Turn a spreadsheet into a batch-generation loop
Sweep every prompt against every checkpoint with one counter
Walk a CSV and a LoRA list off one counter
Log every generation's seed, prompt, and output to a spreadsheet
Skip regenerating what's already on disk
Fetch anything into your workflow without leaving ComfyUI
Pull a voice's details and a sample before you commit to it
Swap in a fallback only when the input is actually blank
Animate a still frame with ByteDance's closed video model
Compose a video from a pile of reference images, clips, and audio
Read a text file's contents into your graph
Fixing the FLOAT-to-INT socket mismatch
Turning a number into text you can actually use
Pull a palette out of any image
One node, single image or batch
Every image in a folder, as a LIST not a batch
Image Overlay
The other half of the socket-type fix
Numbers into text, with an optional format
A guard for the workflow you don't want to run on nothing
Walk a JSON array one item per run
Read a JSON file three ways at once
Pretty-print and minify in one node
How many items, and what's in them
Walk a folder of JSON files by index
Pull up to 10 values out of one blob
Pull one key out, already cast to the type you need
Step through any file type, not just images
Regex-filtered file listing from a folder
The stock loader, plus the filename data ComfyUI won't give you
Load by a path string, not a dropdown
Pull an image straight off the internet
The workhorse for batch runs
One counter, every image × LoRA pair
Four named image slots from a JSON list
Walk every prompt against every LoRA in one counter
A one-file wildcard picker
Read a text file, wrap it, wire it in
Batch through a clips folder like the pack's image loaders
Apply a LoRA by absolute path, no dropdown required
Ten find/replace passes in one node
Merge up to six image batches into one
Glue up to ten text blocks into one prompt
Build a path and filename from reusable pieces
Grow a mask's canvas to match a padded image
Keep a prompt around without retyping it
Full control over the saved filename
Write a string to disk, no fuss
A shrink-only resizer with a real algorithm choice
The math for a capped resize, no resize included
Turn any string into a safe filename
Chain up to five text transforms in one node
Break a string into two pieces on one delimiter
Bridge text output into a real boolean wire
Turn text data into a number your sliders can use
Text to a whole number, for seeds, indexes, and sizes
A plain substring check that returns true/false
A tiny lookup table without a switch chain
A solid RGBA image, made to spec
Google's Veo 3.1 from a reference image, via FAL
The bookkeeping node behind any comparison grid
Stitch a batch into a labeled comparison grid
ComfyUI_Soze
Quality-of-life custom nodes for ComfyUI: batch processing helpers, CSV-driven prompt iteration, file/folder loaders, string and JSON utilities, and integrations for ComfyDeploy, FAL, ElevenLabs, and Azure Blob Storage.
Originally bundled the now-abandoned Comfy_KepListStuff helpers.
Install
Clone into ComfyUI/custom_nodes/:
cd ComfyUI/custom_nodes
git clone https://github.com/SozeInc/ComfyUI_Soze.git
pip install -r ComfyUI_Soze/requirements.txt
Or install via Comfy Registry / ComfyUI Manager.
Configuration
Some nodes need credentials in environment variables (or a .env file in the repo root):
| Variable | Used by |
|---|---|
| CD_API_KEY | ComfyDeploy nodes |
| ELEVENLABS_API_KEY | ElevenLabs voice retriever |
| SOZE_AZURE_STORAGE_CONNECTION_STRING | ComfyDeploy image parameter upload |
| FAL_KEY (or [API] FAL_KEY in config.ini) | FAL nodes (e.g. Veo3.1) |
The repo's .gitignore excludes .env and config.ini. Never commit credentials. If you suspect a key has leaked, rotate it.
Node groups
- CSV —
CSV Reader,CSV Reader X Checkpoint,CSV Reader X Lora,CSV Writer. TheX Checkpoint/X Loravariants iterateindexacross rows × models, so a single counter walks every (row, checkpoint) pair. - Files / folders —
Load Files From Folder,Load Files With Pattern,File Loader,Does File Exist,Save Image With Absolute Filename,Save Text File To Output,Append To Text File. - Images —
Load Image,Load Images From Folder,Load Images From Folder X Lora,Load Image From URL,Load Image From Filepath,Image Batch Process Switch,Image Overlay,Empty Images,Variable Image Builder,Shrink Image,Pad Mask,Get Most Common Image Colors,Multi Image Batch,Soze Image Size With Maximum. - Video —
Append To Video(ffmpeg concat with audio),Load Videos From Folder. - Strings —
Multiline Concatenate,Multi Find And Replace,Special Character Replacer,Text Contains,String Functions,String Splitter,Empty String Replacement,Prompt Cache,Output Filename. - JSON —
JSON Value Parser,JSON Path Extractor,JSON Array Iterator,JSON File Loader,JSON Formatter,JSON Get Array Count,Create Image Batch From JSON Array,Load Images From JSONArray. - Range / XY —
Int/Float Range (Step),Int/Float Range (Num Steps),XY Any,XY Image. - Converters —
Int/Float/Bool/Stringcross-conversion nodes,Boolean Inverter. - Loaders —
Lora File Loader(load by absolute path),Checkpoint File Loader(load by absolute path). - Integrations — ComfyDeploy API queue / cache / download nodes, ElevenLabs voice retriever, FAL Veo3.1 ref-img-to-video.
ffmpeg
Append To Video shells out to ffmpeg; it must be on PATH.
Contributing
Issues and PRs welcome. All respect to the original creators of any subsumed nodes — happy to upstream changes if you'd prefer them in your repo.
License
MIT — see LICENSE.