RedNode Studio is my custom-built creative studio for ComfyUI, designed for painters, designers, and AI artists; from first-time users to experienced workflow builders.

Watch the full video tutorial:

For more advanced text: https://github.com/RedNodeAI/ComfyUI-RedNodeStudio

Update 1.6.4

The Changes

• A tab rail down the side, grouped with icons. Fold it, resize it, put it on either side, and pick a tab preset for the job: Basic, New image, Edit, Paint and more.

• Run tab: Live, Review and Save. Watch the picture form with the pipeline and your VRAM as a chart, look back over this session's runs, and set where everything is filed.

• Paint with AnyPaint. Paint over part of a picture with the normal brush, describe the whole finished picture, and the krea2_anypaint LoRA redraws just that area so it fits the scene. Everything outside the paint stays exactly as it was. About 8 seconds a paint.

• Re-render: turn an illustration into a photograph, as its own tab, a Detailer pass or a Paint engine.

• Tools: Upscale (SeedVR2, VOSR 2.0 or tiled, one picture or a whole folder), Re-angle, Swap and the Converter.

• Hero Creator makes a clean front-on headshot out of any photo for the Identity tab.

• RedNode Shelf: drop pictures on it and drag them anywhere, with five stacks.

• AI tab: pictures to words on a page of its own, and Image to Text as a Detailer pass.

Important fixes

• The Moodboard works properly again. With the Studio preset on custom, the Moodboard and Identity dials were being ignored and every run fell back to Balanced. If the Moodboard felt like it did nothing, this was why.

• Re-render no longer fills the graphics card run after run, and a prompt imported from a picture no longer renders empty.

Handwritten Note

Hi Everyone!

I've been AI painting from November 2023 and been artist for over 15 years. Daily I love the painting process of using AI. I've also hated learning comfy for years, it took me thousands of hours to understand it. The worst thing about the dam software is the look, feel and experience. I wanted to PAINT, not stare at wires and understand how to hook different VAE's into different slots. I want the BULLSHIT of Comfy to be hidden behind TOOLS that do jobs you want. That was the reason I built this.

I do understand that a lot of users want to just press 1 button and go, I prefer to have full control of my images and outputs. So if you are a 1 button person you CAN do that here but its not built for it. Either way I am CHANGING it all the time and now I am making videos about it! So check out the channel and leave a like if you can.

⚡ Install in two clicks (Windows)

The zip has a RedNode Setup folder. Copy its two .bat files into your ComfyUI folder (next to run_nvidia_gpu.bat, or the desktop app's base folder), close ComfyUI, run RedNode_Install_Node_Packs.bat, then RedNode_Download_Models.bat and press Enter for the recommended models. Start ComfyUI, open Templates, RedNode Studio, RedNodeStudio_V1.6.

Not on Windows? Search RedNode Studio in ComfyUI Manager, load RedNode Start Here Install from the templates so Manager can install the other packs, then load RedNodeStudio_V1.6. Every model file and where it goes is on the GitHub page.

The Highlights

Anime to real. Drop an illustration on Re-render and it comes back as a photograph: the same character, pose, outfit and colours, made real with the Anything2Real LoRA.

Watch every render. The Run tab shows the picture forming, every stage with its time, and your VRAM live, with this session's runs down the side. Set a VRAM limit to your card and the run holds models to stay under it.

Paint what you want changed. Brush a mask straight on the picture and generate. With AnyPaint on, a new jacket or a new object fits the scene instead of looking pasted in, and nothing outside the paint changes.

Direct the shot. Drag the camera and the subject on a top view, set height, lens and aim, place people, objects and lights, and the stage writes the camera prompt and drives the camera LoRAs.

Many models, one workspace. Rigs are model setups you switch between: Krea 2, Qwen Image 2.1, Z-Image, SDXL / Illustrious / Pony or your own nodes. Draft on one, refine on another.

Keep a face. Drop a face on Krea 2 Identity and it carries into a new scene, with a Moodboard for the style beside it.

See the whole run before you start. The Workspace Manager lays the run out in order, opens any part in one click, and checks every node pack and model your settings need.

📦 Requirements

ComfyUI 0.27.0 or newer (0.37.0 for Qwen Image 2.1 rigs), Python 3.10 or newer, no pip dependencies.

The core pack downloads nothing and runs no program, and 1.6.4 passed the Comfy Registry's automated scan.

The .bat files are separate helpers you run yourself, and both are plain text you can read first. (put them in your comfy ui install folder)

I HIGHLY RECOMMEND YOU USE THE .BAT FILES. Without them you may take a while looking for the models required to run the tools.

The main files, all from Comfy-Org/Krea-2 on Hugging Face:

krea2_turbo_int8_convrot.safetensors (models/diffusion_models), qwen3vl_4b_fp8_scaled.safetensors (models/text_encoders),

qwen_image_vae.safetensors (models/vae).

Plus the Krea 2 Identity Edit LoRA by lbouaraba for faces, and the krea2_anypaint LoRA by yijunwang2 for AnyPaint. The download .bat fetches all of these; the full list, optional extras included, is on the GitHub page.

Also available for Forge Neo: the Krea 2 moodboard and identity-edit toolkit.