TBG_Enhanced Tiled Upscaler & Refiner FLUX PRO

Welcome to TBG Enhanced Tiled Upscaler and Refiner Pro! We at TBG Think. Build. Generate. AI upscaling & image enrichment are excited to make our TBG Enhanced Tiled Upscaler and Refiner Pro available to you for free non-commercial use. We believe in empowering creators and innovators, which is why anything you create or generate using our software is completely yours to use however you wish—including for commercial purposes!

What This Means for You: Your creations are truly yours: Any images, outputs, or content you generate with our software can be used freely for personal projects, commercial ventures, selling, publishing, or any other purpose you have in mind. Free to use for personal and non-commercial purposes: You can use our Community Edition without any cost for personal projects, research, education, or hobby use.

Important Limitations to Remember:

• only for Comfyui with version Python 3.12 While we want you to have maximum freedom with your creations, there are some important boundaries:
• The software itself (not your outputs) is for non-commercial use only
• You cannot modify, sell, or redistribute the software without our permission
• The Pro Version requires API access authorization
• This is alpha software, so use it at your own risk

Check our license agreement.

Alpha Testing of PRO Features Now Available

We’re excited to announce that alpha testing of the PRO version is now live for our Patreon supporters and free community members!

Get early access by joining us at: https://www.patreon.com/TB_LAAR

You must be a TB_LAAR Patreon supporter or free member to get an API key which you can get here: https://api.ylab.es/login.php

TBG Enhanced Tiled Upscaler and Refiner Pro is an advanced, modular enhancement suite for tiled image generation and refinement in ComfyUI. It introduces neuro generative tile fusion, interactive tile-based editing, and multi-path processing pipelines designed for extreme resolution workflows ut to but not limitet to 100MP, high fidelity, and adaptive post-generation control.

1. New Fusion Techniques

TBG Enhanced Tiled Upscaler and Refiner Pro introduces tree next-generation tile fusion processes that go far beyond traditional blending:

Smart Merge CE

Choose between adaptive blending strategies compatible with common upscalers like ESRGAN, SwinIR, and others. It intelligently handles tile overlaps, avoiding ghosting and seams.

PRO Tile Diffusion

A novel approach where each tile is generated while considering surrounding tile information — perfect for mid-denoise levels. This enables:

• Context-aware detail consistency
• Smooth transitions across tile borders
• Better handling of textures and patterns

PRO Neuro-Generative Tile Fusion (NGTF)

An advanced generative system that remembers newly generated surroundings and adapts subsequent sampling steps accordingly. This makes high-denoise tile refinement possible while maintaining:

• Global consistency
• Sharp, coherent details
• Memory of contextual relationships across tiles

2. Design Suite

TBG_Enhanced is not only about quality — it's about flexible, editable workflows over time:

One-Tile Preview & Smart Presets

Quickly generate single-tile previews to fine-tune the right settings before running the full job. Presets adapt intelligently to image dimensions and desired resolution.

Post-Sampling Tile Correction

You can resample only the tiles you don’t like — no need to regenerate the full image. This allows:

• Precision editing
• Fixing small errors without full reprocessing

Resume Tile Refinement

Modify or refine individual tiles days later while keeping the original input image and final result. The system supports:

• Saved tile maps and settings
• Re-injection of noise and conditioning
• Fully restorable editing workflows

3. Pipeline Structure

TBG_Enhanced is powered by a flexible pipeline architecture built around tile-aware processing paths:

TBG Tile Prompter Pipe

Access prompt and denoise settings per tile. Enables:

• Per-region storytelling
• Adaptive text-to-image behavior
• Denoising strength by tile

TBG Tile Enrichment Pipe

Control multiple sampling and model-level features:

• Model-side: Use DemonTools for deep model manipulation
• Sampler-side: Inject custom noise at specific steps, or apply sigma curves to selected steps
• Sampler-internal: Enable per-step sampler-side noise injection
• Built-in noise reduction
• Optional tile up/downscaling during sampling

ControlNet Pipe

Tiled generation now supports unlimited ControlNet inputs per tile, unlocking:

• High-resolution conditioning
• Fine-grained control for large images
• Targeted structure, depth, edge, pose, or segmentation maps for each region

Getting Started

1. Clone into your ComfyUI custom nodes directory:

   https://github.com/Ltamann/ComfyUI-TBG-ETUR/tree/main

2. Restart ComfyUI. Nodes will appear under the TBG_Enhanced category.

3. (Optional) Place the flux_pro_refiner.safetensors model into your models/ folder if using FLUX PRO.

Status and Roadmap

• Fusion Techniques: Smart Merge, Tile Diffusion, NGTF
• Post-editing Tools: One-tile preview, correction, resume-refine
• Pipelines: Prompt / Enrichment / ControlNet Pipes
• Upcoming: Full integration with custom samplers and DreamEdit compatibility

Instructions for Installing TBG_upscaler_Alfa_1.02p for ComfyUI

1. Copy the folder 'TBG_upscaler_Alfa_1.02p' into your ComfyUI custom_nodes directory: ..\ComfyUI\custom_nodes\

2. Open your ComfyUI Python environment and run the following command: ..\ComfyUI\custom_nodes\TBG_upscaler_Alfa_1.02p\python -m install.py

3. Get early access to PRO feachers by joining us at: https://www.patreon.com/TB_LAAR

   You must be a TB_LAAR Patreon supporter or free member to get an API key which you can get here: https://api.ylab.es/login.php

   You can use your API key in two ways: Paste it directly into the TBG Tiler node

   • OR set an environment variable named: TBG_ETUR_API_KEY This will install the required packages from requirements.txt and download the model file: depth_anything_v2_vitl.pth

---

Alternative Manual Installation:

1. Download the model file manually from: https://huggingface.co/depth-anything/Depth-Anything-V2-Large/resolve/cbbb86a30ce19b5684b7a05155dc7e6cbc7685b9/depth_anything_v2_vitl.pth

2. Copy the downloaded file to the following folder: ..\ComfyUI\custom_nodes\TBG_upscaler_Alfa_1.02p\py\vendor\comfyui_controlnet_aux\ckpts\depth-anything\Depth-Anything-V2-Large

3. Install the required Python packages by running: ..\ComfyUI\custom_nodes\TBG_upscaler_Alfa_1.02p\python -m pip install -r requirements.txt

---

Important: To use PRO features, you must stay connected to the internet. If you're offline, you can still use all the Community Edition features for free.

Recommended Workflow: Do your setup and testing with PRO features turned OFF, and only enable them for the final steps of your workflow.

Thank you for your support and happy tiling!

Just to clarifying the Originality of the TBG Tiled Upscaler CE and PRO Node, this node is not a mere bundle of third-party features—it embodies a completely new tiling concept developed, refined, and rigorously tested over countless hours of engineering and experimentation. Any external code it leverages is strictly minimal and used only where it made sense to avoid reinventing low-level utilities. Here’s exactly what was borrowed and why:

• USDS Crop-Conditioning Node Used solely for its crop-conditioning routine to generate per-tile conditioning masks. All integration and per-tile looping logic around it was handcrafted. https://github.com/ssitu/ComfyUI_UltimateSDUpscale
• LLM Unload Model Function Incorporated to correctly unload large language models between tile runs, preventing memory bloat. The orchestration and tile-level invocation logic are original to this node. https://github.com/willblaschko/ComfyUI-Unload-Models
• MaraScott’s Tile Promter With MaraScott’s explicit permission, their loader and setup code served as a starting scaffold. Every tiling, scheduling, and upscaling mechanism was rewritten or heavily refactored. https://github.com/MaraScott/ComfyUI_MaraScott_Nodes
• KJ Nodes Color-Match Function Employed for its robust color-matching algorithm. Coded into the different tile to tile blendings and not as an one time final process so it had to stay inside the code. https://github.com/kijai/ComfyUI-KJNodes
• ImpactPack Segment Conversion Utilized only to convert internal segment formats to the mask_to_seg format that any ImpactPack node can handel. All tile segmentation, merging, and fusion logic is proprietary. https://github.com/ltdrdata/ComfyUI-Impact-Pack
• EnrichmentPipe Extension - FluxTopas RectifyFlowInversion Brought in for its flow inversion routine; the per-tile applicator and boundary handling are homegrown. https://github.com/logtd/ComfyUI-Fluxtapoz
• EnrichmentPipe Extension - DeamonToles Detail Daemon Node Integrated only for its detail_daemon_node implementation; And names if is. https://github.com/daemontools/daemontools
• ControlNet-Aux Code Borrowed snippets to enable a per-tile ControlNet branch. Crucially, the preprocessor and ControlNet model call within the tile loop are original and make per-tile CNet a much better experience. https://github.com/Fannovel16/comfyui_controlnet_aux