ComfyUI-FunPack

A set of ComfyUI nodes for experimenting with video generation workflows based on WAN, HunyuanVideo, LTX, and similar models.

Updates in 2.6.0

Added FunPack Studio - a single node that combines Refinement Key management, Scene Builder, LoRA loading, Refiner V2, Conditioning Adjust, and sampler configuration into one interface. The node face shows only the rating widget and an Open Studio button. All settings live in a tabbed popup editor.

Popup tabs: Session (refinement key, scene builder mode), Scene (prompt composer, phrase bank), Refiner (all V2 settings, negative prompt, feedback, intent), Advisor (internal HuggingFace LLM), LoRA (session weight suggestions + loading with per-block support), Sampler (Hybrid Euler 2S / Distilled Flow / KSampler for high and low pass, sigma schedules as comma-separated floats), Adjustments (phrase-level conditioning shifts with session phrase bank).

Negative prompt is encoded internally via CLIP when no pre-encoded conditioning is connected, removing the need for an external CLIPTextEncode node in the negative path.

The full LoRA pipeline runs inside Studio: session weight suggestions are read first, LoRAs are applied to model and CLIP, and the Refiner direction patch is applied on top. The model output carries everything.

Two sampler outputs (high pass and low pass) sit between the seed and loss graph outputs. Wire sampler and sigmas directly to your LTX or WAN sampling nodes.

Studio popup remembers the last active tab per node across page refreshes. Field values auto-save 600ms after typing stops.

Also added as standalone nodes: FunPack Conditioning Adjust (phrase-level conditioning adjustment with CLIP-encoded directions, popup editor with session phrase bank).

Fixed FunPackAdvisorLLM crashing with newer transformers versions (apply_chat_template returning BatchEncoding instead of a plain tensor). Fixed advisor producing empty or garbled output for Qwen3 and other thinking models (thinking tokens stripped correctly, attention mask now explicitly provided). Fixed conditioning adjustments not applying for LTX/Gemma3 workflows (was reading pooled_output which is None for T5 encoders, now uses sequence mean). Fixed prompt_repair=False not preventing phrase injection from perfect-repair memory and emphasis. Removed the advisor gate that blocked repair for Perfect-rated outputs.

Updates in 2.5.1

Added FunPack Advisor LLM node - loads any HuggingFace CausalLM (sharded or single-file) as an advisor for Refiner V2 or as a drop-in replacement for the built-in TextGenerate node. Set a HuggingFace repo ID or local path, pick dtype, connect the output to advisor_clip. Model is cached after first load. Compatible with skip_template, min_p, presence_penalty, and any model architecture that AutoModelForCausalLM.from_pretrained supports.

Advisor prompt format rewritten to plain natural language so enhancement-type models (Sulphur, Qwen prompt enhancers) produce useful output rather than ignoring structured field-value instructions. System prompt reduced to one sentence.

Training info Adaptation section now shows direction memory in plain language: run count per slot, magnitude, whether each axis is in direction mode or lerp fallback, and what was applied this run. Model patch status shows which directions were injected into cross-attention and which phrase texts are being emphasized.

Fixed several bugs: advisor returning None when tokenization succeeded (generate was inside the wrong branch), session reset not clearing newer memory fields, system prompt bleeding into advisor output, and repetition loops from the 1.05 penalty being too weak for 8B+ models.

Added persistent cross-run encode cache so phrase encodings are not recomputed every run.

Updates in 2.5.0

Added a two-pass CLIP text-generation advisor to FunPack Video Refiner V2. Pass 1 analyses what specifically needs to change in the suggested prompt. Pass 2 applies those findings. The advisor uses a structured input format — ORIGINAL_USER_INTENT, LAST_PROMPT, RATING, and OPTIONAL_NOTE — so the model knows exactly what failed and what to fix. Token budget: 1200 tokens for analysis, 1600 for repair.

Added feedback_prompt input. When connected, the user's natural-language description of what was wrong is placed first in both advisor passes and the system follows it exactly, overriding axis-based repair logic and validation guards.

Added Prompt only execution mode: all prompt shaping runs normally but conditioning vectors pass through unchanged. Added prompt_repair boolean to disable the rule-based phrase injection from memory when not enough context has been built yet.

Added encoded_prompts STRING output. When the advisor ran it shows up to four sections: Positive prompt (what was encoded), Advisor suggestion (applied/rejected) (the advisor's generated prompt), Advisor analysis (diagnostic from the analysis pass), and Pre-advisor prompt (the prompt before the advisor rewrote it).

Added eta_final parameter to FunPack Hybrid Euler 2S Sampler. When set below eta, ancestral noise decays toward this value as sigma approaches the quality phase boundary. The early phase now uses order-2 denoised extrapolation (Adams-Bashforth 2-step) at zero extra model-call cost, and the quality phase uses progressive correction blending to reduce the number of expensive 2S evaluations.

Removed negative_prompt input and modified_negative conditioning output from Refiner V2. Negative conditioning has no effect at CFG=1.0 and was adding a redundant generation call to every advisor run.

Fixed greedy decoding in the advisor: do_sample was False, causing the model to always produce its highest-probability default output. Fixed Only prompt mode running two generation calls per invocation (positive + the now-removed negative advisor). Fixed encoded_prompts always showing only the positive prompt regardless of advisor activity. Also fixed four logic bugs: intent example lookup reading a nonexistent field, streak updates contaminating conditioning strength in Prompt only mode, unfiltered global phrase memory injecting unrelated vocabulary, and the first-run rating label being forwarded when no previous output existed.

Updates in 2.4.2

Added Refiner V2 Learning mode. It observes prompts, conditioning, ratings, phrase memory, and diagnostics while passing positive and negative prompt conditioning through unchanged.

Fixed Scene Builder mode handling so the live Mode control stays independent from the selected saved scene, and fixed prompt editing so the cursor can move past a final phrase chip with the mouse or right arrow key.

Updates in 2.4.1

Improved Scene Builder database editing: phrase rows now expose the full phrase as a hover hint, and double-click editing opens a wider multiline editor with explicit OK/Cancel buttons.

Updates in 2.4.0

Added FunPack Scene Builder, a replacement for FunPack Template Manager. It stores named scenes from editable positive/negative prompt text, takes prompt/intent text through connection-only inputs, passes the current LoRA stack through unchanged, and can auto-apply a scene when its name or alias appears in an intent prompt.

Added Scene Builder Learning mode for changing-prompt generations: it collects phrase memory inside the selected refinement key while passing the connected prompt data through unchanged, and Refiner resets preserve that Scene Builder memory.

Redesigned Scene Builder as a compact button-driven node with centered editor menus for scene name, mode, aliases, Positive Prompt, Negative Prompt, and Database controls. First use asks for a scene name before editing, prompt editors preserve freely typed text while phrase chips can be clicked or dragged in from the database, used chips are highlighted in the prompt editor, database words can be double-clicked for inline editing, connected prompts now teach useful words as well as phrase chunks, and wildcard random choice is a simple checkbox.

Added searchable LoRA selection to FunPack Apply LoRA Weights while keeping the compact row UI and serialized LoRA stack format.

Added advisory image/CLIP Vision context and repaired negative conditioning to FunPack Video Refiner V2. Vision inputs are stored as metadata and diagnostics only; they are not blended into positive conditioning.

Added an opt-in experimental early velocity bias mode to FunPack Hybrid Euler 2S Sampler for capturing/applying averaged early denoise directions around normalized sigma 0.9 and 0.8.

Updates in 2.3.x

Refiner V2 now supports pre-encoded conditioning workflows, original-intent alignment memory, improved prompt repair scoping, and a Wrong appearance rating for outputs polluted by remembered appearance concepts. Added FunPack Refinement Key Loader.

Dev Branch

The dev branch is intended for testing unfinished changes, implementing new logic and basically, flipping everything just because I can. It can be broken, renamed, or changed without warning.

Use only main if you want the most stable version of this node pack. Bug reports based on dev version will be ignored.

Installation

FunPack is available on Comfy Registry and can be installed in any of these ways:

1. With comfy-cli: comfy node install ComfyUI-FunPack
2. With git, inside your ComfyUI/custom_nodes directory: git clone https://github.com/digital-garbage/ComfyUI-FunPack
3. With ComfyUI-Manager: Open Custom Nodes Manager, search for ComfyUI-FunPack, and click Install.

Dependencies

FunPack includes a requirements.txt file for its Python dependencies.

Install them with:

pip install -r requirements.txt

FunPack uses your existing ComfyUI/PyTorch install. The expected baseline is transformers >= 5.0.0

hpsv3 is optional and only used by the FunPack StoryMem Keyframe Extractor quality filter, so it is not installed by default.

Install it manually only if you need that feature:

pip3 install hpsv3 --no-build-isolation

Important Note About hpsv3

Installing hpsv3 can break Prompt Enhancer and Story Writer, because hpsv3 depends on a transformers version that conflicts with the version those LLM-based nodes require.

FunPack's LLM nodes require transformers >= 5.0. The version required for hpsv3 is strictly transformers==4.45.2. Installing any version different from it will result in broken quality detector.

If you install hpsv3, use --no-build-isolation. Optionally, specify the exact version - pip install transformers==4.45.2 --no-build-isolation.

Documentation

Per-node documentation is available in the docs folder.

Start with:

• docs/FunPackVideoRefinerV2.md for FunPack Video Refiner V2
• docs/FunPackVideoRefinerV2QuickGuide.md for a short Discord-friendly Refiner V2 guide
• docs/FunPackLoraWorkflow.md for the LoRA/refiner helper workflow
• docs/FunPackSceneBuilder.md for scene preset workflows

Version history is available in CHANGELOG.md.

Feedback

If you have suggestions, questions, or ideas for new nodes, feel free to open an issue or submit a pull request.

Thank you

I want to say thanks to teams behind OpenAI (ChatGPT/Codex), xAI (Grok), DeepSeek, Anthropic (Claude) and Google (Gemini) for all the help with coding and transforming my ideas into something working in real UI. Thanks to all the testers and users who regularly use FunPack in their workflows, request features and report bugs. Without all of you, this project would've been just wet dreams of a wannabe coder begging on Discord for someone to add nodes he wants. Seriously, you are cool. I love you all. <3