Boyonodes

BoyoAssetGrabberAdvanced

File: boyoassetgrabberadvanced.py | Category: Boyo/Utility Same as Simple but with explicit path override inputs for custom nodes directory, models directory, etc. Use when ComfyUI is installed in a non-standard location.

Boyo/Asset Management · 0 inputs · 4 parameters · 1 output

BoyoAssetGrabberSimple

File: boyoassetgrabbersimple.py | Category: Boyo/Utility Reads a JSON manifest and automatically installs custom nodes (via git clone), Python packages (via pip), and downloads model files. Detects ComfyUI's base path automatically. Inputs - manifestjson STRING — JSON config defining assets to install - execute BOOLEAN — dry-run when false Outputs - log STRING — installation report

Boyo/Asset Management · 0 inputs · 2 parameters · 1 output

BoyoAudioDurationAnalyzer

File: boyoaudiodurationanalyzer.py | Category: Boyo/Audio/Analysis Extracts the precise duration from a ComfyUI audio tensor. Handles both (batch, channels, samples) and (batch, samples, channels) layouts automatically. Inputs - audio — ComfyUI AUDIO tensor Outputs - durationseconds FLOAT - samplerate INT - info STRING — human-readable summary

Boyo/Audio/Analysis · 1 input · 0 parameters · 2 outputs

BoyoAudioEval

File: BoyoAudioEval.py | Category: VideoUtils Reads an audio file from disk and converts its duration to a frame count. Useful for sizing video generation to match audio length before any audio tensor is loaded into the workflow. Inputs - audiopath STRING — absolute or relative path to a WAV file - fps FLOAT — target frames per second Outputs - framecount INT - videolengthseconds FLOAT - metadata STRING — human-readable summary

VideoUtils · 0 inputs · 2 parameters · 3 outputs

BoyoAudioPadder

Add silence before/after audio with intelligent feedback and auto-centering. Automatically adjusts sample rate for proper ComfyUI playback compatibility. Inputs: - audio (AUDIO) - Input audio to pad - prepadseconds (FLOAT) - Silence before audio (0.0-300.0s) - postpadseconds (FLOAT) - Silence after audio (0.0-300.0s) - targetduration (FLOAT, optional) - Target total duration for comparison - autocenter (BOOLEAN, optional) - Automatically center audio in target duration Outputs: - paddedaudio (AUDIO) - Audio with padding applied (12kHz for proper playback speed) - totalduration (FLOAT) - Final duration after padding - statustext (STRING) - Helpful feedback and suggestions Example Status Messages: - 📊 Audio: 3.2s + Padding: 7.3s = Total: 10.5s ✅ Perfect match! - 📊 Audio: 5.1s + Padding: 2.4s = Total: 7.5s ⚠️ 3.0s shorter than target 💡 Try autocenter or 1.5s each

Boyo/Audio/Processing · 1 input · 4 parameters · 3 outputs

BoyoChainBastardLoops

End-of-chain bastard loops that pull the entire workflow

Boyo/Bastardloops · 2 inputs · 0 parameters · 2 outputs

BoyoChatterboxTurboGenerate

File: boyochatterboxturbogenerate.py | Category: Boyo/Audio/TTS Generate speech from text. Supports emotion tags inline in the transcript and optional voice cloning from a reference audio clip. Inputs - model — handle from Loader node - text — transcript; embed emotion tags anywhere e.g. Hello [chuckle] how are you? - referenceaudio (optional) — 5+ second audio clip to clone the voice from - exaggeration — emotion intensity multiplier - cfgweight — classifier-free guidance weight - seed Outputs - audio — ComfyUI AUDIO tensor at 24 kHz Supported emotion tags: [clear throat] [sigh] [shush] [cough] [groan] [sniff] [gasp] [chuckle] [laugh]

Boyo/Audio/TTS · 2 inputs · 9 parameters · 1 output

BoyoChatterboxTurboLoader

File: boyochatterboxturboloader.py | Category: Boyo/Audio/TTS Loads the Chatterbox Turbo text-to-speech model into memory. Run this once and feed the model handle into the generate node. Inputs - device — cuda or cpu Outputs - model — Chatterbox model handle Notes: Install with pip install chatterbox-tts --no-deps and pip install resemble-perth --no-deps. The --no-deps flags are required to avoid conflicts with ComfyUI's own dependency stack.

Boyo/Audio/TTS · 0 inputs · 1 parameter · 1 output

BoyoCompare

Comparison operations

Boyo/Logic/Math · 1 output

BoyoExecutionBarrier

More aggressive execution barrier that completely isolates execution contexts

Boyo/Utils · 1 input · 2 parameters · 1 output

BoyoForLoopEnd

Exact replication of EasyUse forLoopEnd with reset capability

Boyo/Logic/For Loop · 1 input · 2 parameters

BoyoForLoopStart

Exact replication of EasyUse forLoopStart with reset capability

Boyo/Logic/For Loop · 0 inputs · 3 parameters

BoyoFrameCounter

File: boyoframecounter.py | Category: Boyo/Video Computes framesprocessed = (counter × chunksize) + offset. Feed the loop counter and your chunk size to get the correct framesprocessed value for nodes like VHS Video Combine in looped video workflows. Inputs - counter INT — current loop iteration - chunksize INT — frames per iteration (e.g. 89) - offset INT (optional, default 0) Outputs - framesprocessed INT

Boyo/Video · 0 inputs · 3 parameters · 1 output

BoyoFramePackLoRA

Loads LoRA weights for Hunyuan Video transformer

BoyoNodes · 1 input · 2 parameters · 1 output

BoyoImageCrop

File: boyoimagecrop.py | Category: Boyo Tile-crops a large image into overlapping patches and saves them to a directory. Useful for building tiled datasets from high-resolution source material. Inputs - image IMAGE - cropwidth / cropheight INT — patch size - overlap INT — pixel overlap between adjacent patches - outputpath STRING — destination directory Outputs - status STRING — summary of patches written

Boyo · 1 input · 4 parameters · 1 output

BoyoImageGrab

File: boyoimagegrab.py | Category: Boyo/Image/Editing Monitors a directory and loads the most recently modified image. Designed for iterative semantic editing chains where the output of one pass becomes the input of the next automatically. Inputs - directory STRING - autorefresh BOOLEAN - fileextension — filter by type Outputs - image IMAGE - filename STRING

Boyo/Loaders · 0 inputs · 4 parameters · 4 outputs

BoyoIncontextSaver

Compact description unavailable for this node.

Boyonodes · 2 inputs · 1 parameter · 0 outputs

BoyoIPAdapterZImageLoader

Compact description unavailable for this node.

Boyolatent

Compact description unavailable for this node.

Boyonodes" # Changed From "Latent" To "Boyonodes · 0 inputs · 4 parameters · 1 output

BoyoLatentCacheUpdater

File: boyolatentcacheupdater.py | Category: Boyo/Latent Writes a latent tensor into the Boyo loop cache keyed by a string ID, then passes the tensor through unchanged. Avoids graph cycles by separating the write from the next iteration's read (handled by Boyo Latent Switch). Inputs - latent LATENT - cachekey STRING — must match the key used in Latent Switch - counter INT (optional) Outputs - latent LATENT — unchanged pass-through

Boyo/Latent · 1 input · 2 parameters · 1 output

BoyoLatentPassthrough

Simple latent passthrough that forces a clean execution boundary. Specifically designed to break state contamination between WAN samplers in loops.

Boyo/Utils · 1 input · 2 parameters · 1 output

BoyoLatentSwitch

File: boyolatentswitch.py | Category: Boyo/Latent On iteration 0 outputs startlatent; on all subsequent iterations reads the cached latent written by Boyo Latent Cache Updater. This is the clean way to feed the previous frame's latent back into a sampler without a cycle. Inputs - counter INT - startlatent LATENT — used on first pass - cachekey STRING — must match Cache Updater key - nextlatent LATENT (optional) Outputs - latent LATENT

Boyo/Latent · 2 inputs · 2 parameters · 1 output

BoyoLoadImageBatch

Load images one-at-a-time from a subfolder inside ComfyUI's input directory. The subfolder is chosen from a dropdown — no typing paths — which keeps things reliable on Linux / RunPod deployments. Behaviour is otherwise identical to WAS Load Image Batch: • single_image – load by explicit index • incremental – advance through the folder each execution • random – pick a random image each execution

Boyo Nodes/IO · 2 outputs

BoyoLoadImageList

Compact description unavailable for this node.

Boyonodes · 0 inputs · 5 parameters · 2 outputs

BoyoLoadVideoDirectory

File: BoyoLoadVideoDirectory.py | Category: Boyo/Video Loads every video file in a specified directory as an image batch. Supports common formats (mp4, mov, avi, mkv, webm). Inputs - directory STRING - imageloadcap INT — max frames to load - skipfirstimages INT Outputs - image IMAGE - mask MASK - int — frame count

BoyoNodes/Video · 2 inputs · 10 parameters · 5 outputs

BoyoLoopCollector

Collects and aggregates images from bastard loop iterations

Boyo/Bastardloops · 1 input · 5 parameters · 2 outputs

BoyoLoopCounter

Simple counter that can be reset by the BoyoLoopReset node Alternative to complex loop nodes when you just need a basic counter

Boyo/Logic

BoyoLoopImageSaver

Saves individual images from loop iterations

Boyo/Bastardloops · 1 input · 3 parameters · 1 output

BoyoLoopReset

File: boyoloopreset.py | Category: Boyo/Loops Resets named loop counters to zero when a trigger signal arrives. Use at the end of a workflow to restart a bastard loop from the beginning. Inputs - trigger ANY — completion signal (e.g. from a save node) - resetmode — immediate or delayed - loopid1–3 STRING (optional) — named loop IDs to reset Outputs - trigger ANY — pass-through

Boyo/Logic

BoyoLoRAConfigInspector

File: boyoloraconfiginspector.py | Category: Boyo/LoRA Reads a config file and prints a detailed breakdown: paired vs single LoRA type, all prompt strings, file existence checks, and usage recommendations. Connect before loading to debug config issues. Inputs - configfile STRING Outputs - report STRING — formatted inspection report

Boyo/LoRA Tools · 3 outputs

BoyoLoRAConfigProcessor

File: boyoloraconfigprocessor.py | Category: Boyo/LoRA The logic half of a split architecture. Reads a config and applies the chosen prompt mode and strategy, outputting resolved paths and prompt text. Pair with Boyo LoRA Path Forwarder. Inputs - promptmode — First Only, Cycle Through, Random - config1/2/3 (optional) - strategy1/2/3 - prependtext / appendtext STRING - seed INT Outputs - Resolved LoRA paths and prompt strings

Boyo/LoRA Tools · 7 outputs

BoyoLorainforsender

Simple LoRA selector that outputs just the selected LoRA filename as a string. Stripped down version for index switcher feeding.

Boyo/LoRA Tools · 1 output

BoyoLoRAJSONBuilder

File: boyolorajsonbuilder.py | Category: Boyo/LoRA Creates and saves LoRA configuration JSON files. Each config can hold a paired high-noise and low-noise LoRA path plus an arbitrary list of trigger prompts. Utility LoRAs (e.g. lightning/turbo) can omit prompts entirely. Inputs - configname STRING - highnoiselora STRING — filename from your loras folder - lownoiselora STRING (optional) - prompts STRING — one prompt per line - savedirectory STRING Outputs - configpath STRING — path to the written JSON file

Boyo/LoRA Tools · 2 outputs

BoyoLoRAPairedLoader

File: boyolorapairedloader.py | Category: Boyo/LoRA Loads up to three LoRA configs at once and resolves prompts from each according to a per-slot strategy. Outputs all six LoRA paths and four combined prompt strings ready for downstream nodes. Inputs - config1/2/3 — config file names (dropdowns from your config directory) - strategy1/2/3 — Mute, Concatenate, or Merge - promptmode — First Only, Cycle Through, or Random - seed INT — used for deterministic random prompt selection Outputs - lorahigh1–3 STRING — high-noise LoRA paths - loralow1–3 STRING — low-noise LoRA paths - prompt1–4 STRING — resolved combined prompts Typical usage: slot 1 = utility LoRA (strategy: Mute), slots 2–3 = character + style (strategy: Concatenate).

Boyo/LoRA Tools

BoyoLoRAPathForwarder

File: boyolorapathforwarder.py | Category: Boyo/LoRA Buffers LoRA filenames and forwards them to standard ComfyUI LoRA loader nodes. Caches the LoRA list at startup to prevent dynamic type-change issues during long batch runs. Inputs - lorahigh / loralow STRING — filenames from Processor or Paired Loader Outputs - lorahigh / loralow STRING — same filenames, now stable for downstream loaders

Boyo/LoRA Tools · 3 inputs · 1 parameter

BoyoMaskToImage

File: boyomasknormalise.py | Category: BoyoNodes/Masks Converts a mask tensor of any standard shape to a 3-channel IMAGE tensor. Written specifically to work around DiffuEraser's broken len(videomask) > 3 check, which incorrectly rejects any video longer than 3 frames. Inputs - mask MASK — accepts [h,w], [b,h,w], [b,h,w,1], or [b,1,h,w] Outputs - image IMAGE — shape [b,h,w,3]

BoyoNodes/Masks · 1 input · 0 parameters · 1 output

BoyoMathInt

Simple math operations for integers

Boyo/Logic/Math · 0 inputs · 3 parameters · 1 output

BoyoOverlapSwitch

File: boyooverlapswitch.py | Category: Boyo/Video Outputs a different overlap value depending on whether this is the first loop iteration or a subsequent one. Prevents the first generated chunk from overlapping nothing. Inputs - counter INT - firstoverlap INT — typically 0 - subsequentoverlap INT — your optimal blend value (default 13) Outputs - overlap INT

Boyo/Video · 0 inputs · 3 parameters · 1 output

BoyoPainterSVI

File: boyopaintersvi.py | Category: conditioning/videomodels Merges PainterI2V motion amplitude conditioning with WanImageToVideoSVIPro context preservation. Intended for samplers 2 and onwards in infinite-length video generation, after the first PainterI2V pass has established motion. Inputs - positive / negative CONDITIONING - length INT — frame count for this chunk - anchorsamples LATENT — latent from the first PainterI2V sampler - motionamplitude FLOAT (1.0–2.0) - motionlatentcount INT - prevsamples LATENT (optional) — previous chunk for context blending Outputs - positive / negative CONDITIONING - latent LATENT

Conditioning/Video Models · 4 inputs · 3 parameters · 3 outputs

BoyoPairedImageSaver

File: boyopairedimagesaver.py | Category: Boyo/Image/Editing Saves an original and an edited image as a sequentially numbered pair. Compatible with ControlNet dataset formats. Use alongside Boyo Image Grab to build an iterative editing pipeline. Inputs - originalimage IMAGE - editedimage IMAGE - outputdirectory STRING - filenameprefix STRING Outputs: none (output node)

Boyonodes · 2 inputs · 2 parameters · 0 outputs

BoyoPairedSaver

Compact description unavailable for this node.

Boyonodes · 2 inputs · 2 parameters · 0 outputs

BoyoPiDAspectRatio

Outputs LDM width/height + PiD width/height for a chosen profile+ratio.

BoyoNodes/PiD · 1 input · 1 parameter · 5 outputs

BoyoPromptInjector

Injects prompts from the loop configuration into the workflow

Boyo/Bastardloops · 1 input · 2 parameters · 1 output

BoyoPromptLoop

File: BoyoPromptLoop.py | Category: Boyo/Loops Iterates through prompts stored in .txt files in a prompts/ subfolder. Three modes: sequential (one per loop pass), random (seeded), or single (fixed index). Inputs - textfile — dropdown of .txt files in prompts/ directory - mode — sequential, random, or single - startseed INT - index INT — used in single mode Outputs - prompt STRING - index INT — current position

Boyo/Bastardloops · 1 output

BoyoPromptRelayEncode

File: boyopromptrelaynodes.py | Category: (conditioning) Port of kijai's ComfyUI-PromptRelay into Boyonodes. Encodes a set of prompts with temporally-varying Gaussian segment masks so different prompts influence different time steps of a video generation. Inputs - clip CLIP - prompts — list of text segments - segmentlengths — duration of each segment as a proportion of total steps Outputs - conditioning CONDITIONING Attribution: Original logic by Jukka Seppänen (kijai), ported with attribution.

3 inputs · 4 parameters

BoyoPromptRelayEncodeTimeline

File: boyopromptrelaynodes.py | Category: (conditioning) Timeline-based variant of Prompt Relay. Accepts explicit step-range definitions per prompt segment instead of proportional lengths, for finer control.

3 inputs · 8 parameters

BoyoPromptRelayLoraGate

File: boyopromptrelaynodes.py | Category: (conditioning) Applies a LoRA selectively to a single temporal segment using the same Gaussian gating mechanism as Prompt Relay. Enables LoRA influence to be confined to a specific portion of the video. Inputs - model MODEL - loraname STRING - strength FLOAT - segmentstart / segmentend FLOAT — 0–1 proportional range Outputs - model MODEL For installation instructions and one-line node summaries see README.md.

1 input · 3 parameters

BoyoQwenVLGrounding

File: boyoqwengrounding.py | Category: (Boyo/Vision implied) Runs Qwen2.5-VL visual grounding to locate objects in an image and draw annotated bounding boxes. Supports 4-bit quantisation for lower VRAM usage. Inputs - image IMAGE - modelcheckpoint — Qwen2.5-VL model path - prompt STRING — grounding query e.g. "the red car" - quantize4bit BOOLEAN Outputs - image IMAGE — annotated with bounding boxes - groundingjson STRING — raw detection results Dependency: pip install qwen-vl-utils

BoyoNodes/QwenVL · 1 input · 10 parameters · 4 outputs

BoyoResolutionCalc

File: BoyoResolutionCalc.py | Category: Boyonodes Takes a base width and a named aspect ratio and returns width and height both rounded to the nearest multiple of 8. Inputs - width INT - aspectratio — 1:1 Square, 4:3, 3:2, 16:9 HD, 21:9, 2:3, 3:4, 9:16 Tiktok-Shorts Outputs - width INT - height INT

Boyonodes · 0 inputs · 2 parameters · 2 outputs

BoyoSaver

Compact description unavailable for this node.

Boyonodes · 2 inputs · 2 parameters · 0 outputs

BoyoStoryboardJsonParser

File: boyostoryboardjsonparser.py | Category: Boyo/Storyboard Alternative parser for storyboard JSON with the same 12-output structure. Use if the primary output node has trouble with a particular model's JSON formatting. Inputs / Outputs: identical to Boyo Storyboard Output.

Boyo/Storyboard · 0 inputs · 1 parameter · 1 output

BoyoStoryboardOutput

File: boyostoryboardoutput.py | Category: Boyo/Storyboard Parses the JSON string returned by an ollama node and splits it into 12 individual prompt outputs for direct connection to image and video generation nodes. Inputs - jsonstring STRING — raw JSON from ollama Outputs - image1 … image6 STRING - video1 … video6 STRING

Boyo/Storyboard · 0 inputs · 7 parameters · 1 output

BoyoStoryboardPrompt

File: boyostoryboardprompt.py | Category: Boyo/Storyboard Constructs a system prompt and user prompt that instructs an ollama model to generate a structured 6-scene storyboard in JSON. Two modes are available: standard (6 image + 6 video prompts) and travelling (6 images + 6 multi-line video sequences for extended content). Inputs - storyoutline STRING - characterdescription STRING - styledescription STRING - mode — Standard or Travelling - additionaldetails STRING (optional) - modeltriggerword STRING — LoRA/video model trigger Outputs - systemprompt STRING - userprompt STRING Recommended model: Qwen 30B A3B Coder Abliterated (via ollama). Avoid Gemma, Meta coding variants, and thinking models.

Boyo/Storyboard · 0 inputs · 6 parameters · 3 outputs

BoyoTiledVAEDecode

Compact description unavailable for this node.

Boyonodes · 2 inputs · 3 parameters · 1 output

BoyoVACEInjector

File: BoyoControl.py | Category: Boyo/Control Injects VACE control data directly into model attributes, bypassing conditioning entirely. Supports standard ComfyUI MODEL or WanVideoWrapper's WANVIDEOMODEL. Inputs - controlimage IMAGE - vacestrength FLOAT (0–2) - vacestartpercent / vaceendpercent FLOAT - numframes INT - model MODEL (optional) - wanvideomodel WANVIDEOMODEL (optional) - vae VAE (optional) Outputs - model MODEL

BoyoResearch · 4 inputs · 4 parameters · 1 output

BoyoVACEViewer

File: BoyoControl.py | Category: Boyo/Control Reads VACE control attributes from a model and displays them as a formatted string for debugging injection results. Inputs - model MODEL Outputs - info STRING

BoyoResearch · 1 input · 0 parameters · 1 output

BoyoVAEDecode

Compact description unavailable for this node.

Boyonodes · 2 inputs · 0 parameters · 1 output

BoyoVideoClipper

File: boyovideoclipper.py | Category: Boyo/Video Clips a video from ComfyUI's input directory to an exact frame count starting from a given time. Uses FFmpeg internally. Useful for preparing dataset clips at a precise resolution and frame rate. Inputs - video — dropdown from input directory - starttime FLOAT — seconds - targetfps INT - requiredframes INT Outputs - IMAGE batch — decoded frames - framecount INT

Boyo/Video · 0 inputs · 4 parameters · 3 outputs

BoyoVideoCutter

File: boyovideocutter.py | Category: Boyo/Video Removes specific frames from an image sequence, typically the overlapping frames at chunk junctions in looped video generation. Pairs directly with Boyo Video Length Calculator. Inputs - images IMAGE — full frame batch - trimpositions STRING — frame indices to remove, e.g. "94,95,96|188,189,190" (pipe separates junction groups) - debugmode BOOLEAN Outputs - trimmedimages IMAGE - cutinfo STRING - framesremoved INT

Boyo/Video · 1 input · 2 parameters · 3 outputs

BoyoVideoLengthCalculator

File: boyovideolengthcalculator.py | Category: Boyo/Video Looks up a hardcoded table to return the total frame count, number of loop iterations, and junction trim positions needed to produce a target video duration. Works in 5-second increments up to 60 seconds. Inputs - targetseconds INT — 5 to 60, step 5 Outputs - totalframes INT - loopsneeded INT - trimpositions STRING — feed directly into Boyo Video Cutter - info STRING

Boyo/Video · 0 inputs · 1 parameter · 4 outputs

BoyoVideoPairedSaver

File: boyovideopairedsaver.py | Category: Boyo/Video Renders an image batch to a video file via FFmpeg and saves a matching .txt prompt file alongside it. Supports optional audio muxing. Useful for building video-text pair datasets. Inputs - images IMAGE — frame batch - enhancedprompt STRING - foldername / filenameprefix STRING - fps FLOAT - codec — libx264, libx265, or av1 - quality — high, medium, or low - audio AUDIO (optional) Outputs: none (output node)

Boyonodes · 2 inputs · 5 parameters · 0 outputs

BoyoVideoSaver

Simple video saver without prompt pairing - just saves the video

Boyonodes · 2 inputs · 5 parameters · 1 output

BoyoVision

Compact description unavailable for this node.

BoyoNodes/QwenVL · 1 input · 8 parameters · 1 output

BoyoVoiceEnhancer

File: boyovoiceenhancer.py | Category: Boyo/Audio/TTS Enhances or converts voice characteristics using the Seed-VC model. Requires the seed-vc submodule and its dependencies. Inputs - audio — source audio tensor - referenceaudio — target voice reference - Various Seed-VC inference parameters (diffusion steps, pitch shift, etc.) Outputs - audio — enhanced/converted audio tensor Setup: git submodule add https://github.com/Plachtaa/seed-vc.git seed-vc && git submodule update --init --recursive

Boyo/Audio/VoiceEnhancement · 2 inputs · 8 parameters · 1 output

BoyoWatermarks

File: BoyoWatermarks.py | Category: BoyoNodes/video Stamps a watermark image onto every frame of a video batch. The watermark is resized to specified dimensions, composited at a chosen corner with adjustable opacity. Inputs - frames IMAGE — video frame batch - watermark IMAGE — single watermark frame - wmwidth / wmheight INT - corner — bottom-right, bottom-left, top-right, top-left - opacity FLOAT (0–1, default 0.85) - padding INT — pixel gap from edge Outputs - frames IMAGE

BoyoNodes/Video · 1 output

BoyoWhileLoopEnd

Exact replication of EasyUse whileLoopEnd with reset capability

Boyo/Logic/While Loop

BoyoWhileLoopStart

Exact replication of EasyUse whileLoopStart with reset capability

Boyo/Logic/While Loop

MandelbrotVideo

Compact description unavailable for this node.

Video · 0 inputs · 3 parameters · 2 outputs

node_id

Compact description unavailable for this node.