Save 4 hours! We auto-setup your workflow! Free!

Drop your workflow.json — we handle every dependency, custom node, and model. Just open the link and run.

Auto-Setup Workflow Json (Free) Now!
ComfyUI > Nodes > ComfyUI-988

ComfyUI Extension: ComfyUI-988

Repo Name

ComfyUI-988

Author
kajan988 (Account age: 0 days)
Nodes
View all nodes(1)
Latest Updated
2026-11-03
Github Stars
0.0K

How to Install ComfyUI-988

Install this extension via the ComfyUI Manager by searching for ComfyUI-988
  • 1. Click the Manager button in the main menu
  • 2. Select Custom Nodes Manager button
  • 3. Enter ComfyUI-988 in the search bar
After installation, click the Restart button to restart ComfyUI. Then, manually refresh your browser to clear the cache and access the updated list of nodes.

Visit ComfyUI Online for ready-to-use ComfyUI environment

  • Free trial available
  • 16GB VRAM to 80GB VRAM GPU machines
  • 400+ preloaded models/nodes
  • Freedom to upload custom models/nodes
  • 200+ ready-to-run workflows
  • 100% private workspace with up to 200GB storage
  • Dedicated Support

Run ComfyUI Online

ComfyUI-988 Description

ComfyUI-988 adds nodes for local LM Studio text and vision responses plus tiled image upscaling: plan and extract tiles, process them, then blend overlapping edges into a seamless image.

ComfyUI-988 Introduction

ComfyUI-988 is a collection of ComfyUI nodes for two kinds of creative work: generating text with local language models and processing images in tiles. Its LM Studio nodes can bring text or image-aware model responses into a workflow. Its image nodes help you plan an upscale, divide the image into manageable pieces, and blend the processed pieces back together.

This can be useful when an image is too large to process comfortably in one pass, or when you want a local model to help produce text within a ComfyUI workflow. The extension also includes a Photopea frontend integration, although the provided information does not describe its controls or workflow.

How ComfyUI-988 Works

Think of a ComfyUI workflow as a set of connected steps. Each node does one job and passes its result to the next node.

For text generation, LM Studio 988 sends a request to a local model through an LM Studio server and receives generated text. If you provide images to a vision-capable model, the model can use those images as part of its response. The node can also separate reasoning text for supported model families and use a system message template to guide how the model responds.

For tiled image processing, the nodes work together:

  1. DaC Algorithm 988 v2 plans upscale dimensions and tile coordinates. It accounts for a minimum scale factor and the overlap needed between neighboring tiles.
  2. Divide Image Select 988 uses those coordinates to extract all tiles or one selected tile.
  3. Your workflow processes the tiles.
  4. Combine Tiles 988 v2 puts them back into one image. It blends overlapping areas with Gaussian-blurred masks to make the joins less noticeable.

The overlap is like leaving a little extra image around the edge of each puzzle piece. That shared area gives the combine step room to blend adjacent pieces instead of placing them edge to edge.

ComfyUI-988 Features

NodeWhat it doesChoices you can make
LM Studio 988Generates text through an LM Studio server using a local language model or vision-language model.You can use multiple images with a vision-capable model, use system message templates, extract reasoning from supported model families, or use a draft model for speculative decoding. The available models depend on your LM Studio setup.
LM Unload 988Optionally unloads language models from LM Studio and/or clears ComfyUI VRAM. It is a passthrough node, so it can sit within a workflow without serving as a text-generation step.Use the unload or VRAM-clearing actions when you need to free resources for later steps.
DaC Algorithm 988 v2Calculates upscale dimensions while maintaining tile-overlap constraints.The minimum scale factor affects how large the planned result must be; tile and overlap constraints affect how that result is divided.
Divide Image Select 988Extracts tiles from the upscaled image using the DaC coordinates.Set tile=0 to return all tiles, or tile=N to return one tile. Selecting one tile can help you inspect or test part of a workflow.
Combine Tiles 988 v2Reassembles processed tiles into one image and blends their overlapping edges.Its result depends on receiving tiles that correspond to the planned layout. The overlap masks help reduce visible joins.
DAC Data Scale 988Multiplies the width, height, tile, and overlap dimensions in DAC_DATA by a chosen value.Use a multiplier when you need the same tile plan at a different scale. The grid arrangement and tile order stay the same.

The extension also contains a Photopea integration in its frontend files. The provided documentation does not specify what actions it adds, so no particular editing behavior is assumed here.

ComfyUI-988 Models

ComfyUI-988 does not provide a fixed list of included models in the available documentation. LM Studio 988 works with local LLMs and vision-language models served through LM Studio.

Choose a language model for text-only tasks. Choose a vision-capable model when you want to include images in the request. Reasoning extraction is described for DeepSeek, Qwen, QwQ, GLM, and GPT-OSS model families; these are supported families, not a list of models bundled with the extension. Draft models are also supported for speculative decoding, but the documentation does not name specific draft models or promise a particular speed improvement.

The DaC image nodes are processing tools, not AI models. They plan, select, scale, and combine tiles around whatever image-processing steps you place between them.

Troubleshooting ComfyUI-988

LM Studio 988 is not producing text

  1. Check that your LM Studio server is running.
  2. Check that the model you intend to use is available through that server.
  3. If your request includes images, confirm that the selected model supports vision.
  4. Try a simple text-only request to narrow down whether the issue concerns the connection, the model, or the image input.

The image tiles do not combine as expected

  1. Confirm that Divide Image Select 988 is using coordinates from the intended DaC plan.
  2. Check whether you selected all tiles with tile=0 or only one tile with tile=N.
  3. Make sure the processed tiles correspond to the same grid and tile order used for the combination step.
  4. If you scaled DAC_DATA, check that the image and tile-processing steps use the corresponding dimensions.

The joined image has noticeable boundaries

Check that your workflow retains the planned overlap between tiles. Combine Tiles 988 v2 uses overlapping regions for its blended joins; a workflow that removes or changes those regions may not give it the expected material to blend.

Frequently asked questions

  • Do I need a vision model to use LM Studio 988? No. A language model can be used for text generation; image inputs require a vision-capable model.
  • What does tile=0 mean? It returns all tiles. A numbered tile selection returns one tile.
  • Does DAC Data Scale 988 rearrange the image grid? No. It scales dimensions while preserving the grid and tile order.
  • Does LM Unload 988 generate text? No. Its documented purpose is to unload an LM Studio model and/or clear ComfyUI VRAM as an optional passthrough step.

Learn More about ComfyUI-988

Start with the included DaC upscale example workflow. Its preview image provides a visual reference for the workflow.

Parts of ComfyUI-988 are adapted from other projects. Their repositories may provide useful background on the underlying tools: EA_LMStudio for the LM Studio-related nodes, ComfyUI_Steudio for the DaC image nodes, and ComfyUI-Photopea for the Photopea integration. These are source-project references; their documentation may not describe ComfyUI-988’s adapted nodes exactly.

ComfyUI-988 Related Nodes

RunComfy
Copyright 2025 RunComfy. All Rights Reserved.

RunComfy is the premier ComfyUI platform, offering ComfyUI online environment and services, along with ComfyUI workflows featuring stunning visuals. RunComfy also provides AI Models, enabling artists to harness the latest AI tools to create incredible art.

ComfyUI-988 detailed guide | ComfyUI