A growing collection of custom ComfyUI nodes for AI video creation, music videos, storyboarding, image generation, video enhancement, face repair, editing, LoRA training, and workflow automation.

The flagship tool is the AI Video Builder: a scene-by-scene production workspace for LTX 2.3, MiniMax H3, and future video engines that brings planning, prompting, media generation, timing, review, and final assembly together inside ComfyUI.

The AI Video Builder shows host RAM and NVIDIA VRAM usage in its top toolbar, refreshing about every two seconds while the Builder is open. The compact display hides automatically when the toolbar is too narrow. GPU readings require the driver's nvidia-smi; RAM remains available when GPU sensors are unsupported.

The optional canvas strip can be enabled under Settings → VRGDG → Resource Monitor. It adds GPU load, temperature, RAM/VRAM bars, optional fan/clock/power details, and an idle-only Clear memory action. Readings come from the machine running ComfyUI.

Add the node named VRGDG AI Video Builder UI to open the Builder.

Use it to:

• 🎵 Build projects from songs, audio, SRT files, lyrics, or manually timed scenes.
• 🧙 Start quickly with the guided Wizard, Storyboard Builder, and Reference Builder.
• 🖼️ Plan and generate scene images with character and location references.
• 🎞️ Create image-to-video, text-to-video, ID-LoRA, and First/Last Frame scenes.
• 🔗 Build chained or independent First/Last Frame sequences for stronger continuity.
• 🗣️ Review lyrics, map singers, plan lip-sync scenes, and mark instrumental or B-roll sections.
• ✨ Repair faces, enhance clips, apply post-processing, and compare results.
• 🎚️ Preview timing on the visual timeline, calibrate beat markers, and stitch the final video.
• 💾 Save, branch, export, import, and continue portable Builder projects.
• 🤖 Use built-in LLM, local LM Studio, API, and Browser AI options where supported.

📖 New here? Start with the full AI Video Builder Guide.

✨ Or chat with a GPT and ask any question about the video builder. HERE

The skills/vrgdg-h3-film folder is a skill for Claude Code. With it, Claude writes, renders, scores, checks and edits a complete short film in your own ComfyUI through the Video Builder, using MiniMax H3 with built-in voices.

Copy that folder into your Claude skills folder (C:\Users\<you>\.claude\skills\ on Windows, ~/.claude/skills/ on Mac or Linux), then ask Claude for a film. Requirements and setup are in the skill's README.

These are some of the most useful tools included in the pack. Many can be used on their own in a normal ComfyUI workflow.

• VRGDG Storyboard Creator with Browser AI — Open This — Create project-aware start/end storyboard images with supported browser image tools.

• Face Fix node set — Detect and track a face, enhance guided anchors, process the crop through LTX, and composite the repaired face back into the source video. Start with the included Face Fix workflow.
• Video Enhance node set — Create guided enhancement anchors, process a full video through LTX, and restore the exact original resolution and frame count.
• Z-Image upscaler workflows — Upscale and refine images with Z-Image using the included workflows for several source pipelines. Browse the Z-Image Upscale workflows.
• Image comparison tools — Compare an original and processed image directly inside ComfyUI.
• Fast Film Grain, Color Match, and Sharpening nodes — Add cinematic grain, match a reference palette, or restore edge detail efficiently across image batches.

• VRGDG LoRA Dataset Creator UI — Build and review captioned datasets for styles, characters, and experimental edit pairs.
• MiniMax H3 and LTX 2.3 and Z-Image LoRA training workflows — Train standard video, audio, audio/video, and Speed LoRAs with the included updated LoRA training workflows.
• VRGDG Musubi-Tuner Installer and Krea 2 tools — Set up supported training environments and use preset-based training, sampling, and comparison tools.
• Preview and grid plot nodes — Compare checkpoints, prompts, strengths, and generated video folders.

• VRGDG VoxCPM2 Voice Clone / TTS — Generate speech from text, design a voice, continue spoken audio, or clone a voice from a reference clip. Start with the included VoxCPM2 Voice Clone / TTS workflow.
• Audio loading, splitting, timing, transcription, and silent-audio helpers.
• Local and API-based LLM prompt tools for structured image and video prompting.
• Image, text, switch, folder, workflow-runner, and batch-processing utilities.
• LUT, color, grain, sharpening, resize, combine, and general video-processing nodes.

Some advanced tools need additional models or external components. The Builder guide explains the supported workflows, required custom nodes, model locations, and optional setup.

1. Install the node pack and restart ComfyUI.
2. Hard refresh the ComfyUI browser page so the latest JavaScript UI files load.
3. Add VRGDG AI Video Builder UI.
4. Create a project and add audio, SRT timing, or scenes.
5. Use the Wizard or Storyboard Builder to plan the project.
6. Generate and approve scene images, render scene videos, then stitch the final video.

1. Open Manager → Install Custom Nodes.
2. Search for vrgamedev, or install from:

https://github.com/vrgamegirl19/comfyui-vrgamedevgirl

3. Restart ComfyUI and hard refresh the browser page.

Clone this repository into ComfyUI/custom_nodes:

git clone https://github.com/vrgamegirl19/comfyui-vrgamedevgirl.git

Then install the Python requirements using the Python environment that runs ComfyUI. For the Windows portable build, run this from ComfyUI_windows_portable:

python_embeded\python.exe -m pip install --upgrade pip setuptools wheel
python_embeded\python.exe -m pip install Cython scikit-build-core
python_embeded\python.exe -m pip install -r ComfyUI\custom_nodes\comfyui-vrgamedevgirl\requirements.txt

The first two commands prepare the build tooling needed by voxcpm and llama-cpp-python, especially on Windows with Python 3.13. llama-cpp-python may also need a working CMake/Ninja and C++ compiler when a compatible prebuilt wheel is unavailable. Python 3.12 is the safer choice for older Windows portable environments.

Restart ComfyUI and hard refresh the browser page after installation.

• The Video Builder guide is the main source for setup, model paths, screenshots, and step-by-step help.
• Start with a short project or a few scenes before committing to a long render.
• Save often, and use Branch Project or Export Shareable Project ZIP before major experiments.
• Optional training, Browser AI, face-repair, and enhancement tools may have their own setup requirements.

Created by VRGameDevGirl ✨

• 💬 Join the Discord community
• ☕ Support VR Game Dev Girl
• 📺 Videos created with these workflows
• 🎓 Walkthroughs and update videos — Some features have been updated since these were recorded, so the current UI may look different.

Covered original code is available under the VRGameDevGirl Community and Commercial License. Copyright (c) 2026 Jean Thompson (VRGameDevGirl).

You may create, sell, and monetize videos and other creative media without paying a software license fee, including paid client work and internal studio production. Model licenses and other third-party rights still apply.

Free community forks and modifications are permitted under the license. Commercializing the covered software itself requires a separate paid written license: this includes selling the code or forks, commercial integrations, paid software access, and commercial hosted generation services.

For commercial licensing, contact Jean Thompson at jeanthompson1984@gmail.com. A usage license does not transfer copyright ownership.

This is a source-available license. Models, ComfyUI, dependencies, and third-party code retain their own terms. Previously granted AGPL and other rights are not revoked. See LICENSE for the full terms and the licensing guide for examples and transition details.