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-Utility-Suite > Image Tile Batch

ComfyUI Node: Image Tile Batch

Class Name

UtilitySuiteImageTileBatch

Category
Utility Suite/Image/Tiling
Author
tom-m-2020 (Account age: 0days)
Extension
ComfyUI-Utility-Suite
Latest Updated
2026-11-12
Github Stars
0.0K

How to Install ComfyUI-Utility-Suite

Install this extension via the ComfyUI Manager by searching for ComfyUI-Utility-Suite
  • 1. Click the Manager button in the main menu
  • 2. Select Custom Nodes Manager button
  • 3. Enter ComfyUI-Utility-Suite 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

Image Tile Batch Description

Splits image batches into uniform, edge-padded tiles with overlap and source-position metadata.

Image Tile Batch:

Image Tile Batch divides an image or batch of images into smaller, consistently sized tiles. This lets you work on separate regions while retaining a record of where each region came from. You can choose fixed-size tiles, limit the number of tiles across each dimension, or request a uniform grid. When a tile extends past an image edge, the node fills the remaining area by repeating the edge pixels, so every tile in the resulting batch has the same dimensions. It also keeps the tile order and placement information needed to associate processed tiles with their original images.

Image Tile Batch Input Parameters:

image

The IMAGE to divide into tiles. You can supply one image or an image batch; the node processes each image using the same tile layout. The image must have the standard four-dimensional IMAGE shape: batch, height, width, and channels. Every dimension must be greater than zero. No default, minimum, or maximum image size is specified in the node interface.

mode

Selects how the tile layout is calculated. Options are fixed_tile, bounded_grid, and uniform_grid; no default is specified. fixed_tile uses the requested tile dimensions and overlap, creating as many rows and columns as needed. bounded_grid limits the row and column counts and may make tiles larger than the requested dimensions to cover the image. uniform_grid uses the requested tile dimensions together with the minimum-overlap and minimum-count settings; the regular overlap and maximum-count settings do not govern its layout.

tile_width

The requested width of each tile, in pixels. In fixed_tile and uniform_grid modes, it is also the resulting tile width. In bounded_grid mode, the resulting width may increase to fit within the column limit. Default: 1024; minimum: 1; maximum: 32768. Smaller tiles can produce more pieces, while larger tiles cover more of the image at once.

tile_height

The requested height of each tile, in pixels. In fixed_tile and uniform_grid modes, it is also the resulting tile height. In bounded_grid mode, the resulting height may increase to fit within the row limit. Default: 1024; minimum: 1; maximum: 32768. Adjust it independently of width when your image or downstream processing calls for a different tile shape.

overlap_x

The requested horizontal overlap between neighboring tiles, in pixels, for fixed_tile and bounded_grid modes. Overlap gives adjacent tiles some image content in common; increasing it generally means more horizontal tiles may be needed. It must pass the layout’s overlap validation relative to the tile width. It is not used to calculate the uniform_grid layout. Default: 128; minimum: 0; maximum: 16384.

overlap_y

The requested vertical overlap between neighboring tiles, in pixels, for fixed_tile and bounded_grid modes. Increasing it gives tiles in adjacent rows more content in common and may require more rows. It must pass the layout’s overlap validation relative to the tile height. It is not used to calculate the uniform_grid layout. Default: 128; minimum: 0; maximum: 16384.

max_columns

The maximum number of tile columns in bounded_grid mode. If the requested tile width and overlap would otherwise produce more columns, the node limits the count and calculates a wider tile to cover the image. The setting does not cap the column count in fixed_tile or uniform_grid mode. Default: 3; minimum: 1; maximum: 256.

max_rows

The maximum number of tile rows in bounded_grid mode. If the requested tile height and overlap would otherwise produce more rows, the node limits the count and calculates a taller tile to cover the image. The setting does not cap the row count in fixed_tile or uniform_grid mode. Default: 3; minimum: 1; maximum: 256.

min_overlap_x

The requested minimum horizontal overlap for uniform_grid mode. Use it when you want neighboring columns to share at least a specified amount of image content under that mode’s layout rules. It does not control the fixed_tile or bounded_grid layout. Default: 128; minimum: 0; maximum: 32768. The selected value must be compatible with the requested tile width and image dimensions for a valid uniform layout.

min_overlap_y

The requested minimum vertical overlap for uniform_grid mode. Use it when you want neighboring rows to share at least a specified amount of image content under that mode’s layout rules. It does not control the fixed_tile or bounded_grid layout. Default: 128; minimum: 0; maximum: 32768. The selected value must be compatible with the requested tile height and image dimensions for a valid uniform layout.

min_columns

The requested minimum number of columns for uniform_grid mode. It influences the uniform layout and does not set the column count in the other modes. Raising it requests more horizontal divisions, provided the image dimensions, tile width, and overlap requirements permit a valid layout. Default: 1; minimum: 1; maximum: 32768.

min_rows

The requested minimum number of rows for uniform_grid mode. It influences the uniform layout and does not set the row count in the other modes. Raising it requests more vertical divisions, provided the image dimensions, tile height, and overlap requirements permit a valid layout. Default: 1; minimum: 1; maximum: 32768.

Image Tile Batch Output Parameters:

tiles

An ordinary IMAGE batch containing the extracted tiles. Every tile has the same width and height. Tiles are ordered by source image first, then by their spatial position in row order, from left to right within each row. Where a tile reaches beyond the source image, its unused area is filled by repeating its last valid edge pixels. This output is the image data you can pass to subsequent image-processing nodes.

layout

A TILE_LAYOUT value describing the original image dimensions, resolved tile dimensions, grid size, selected mode, requested settings, and each tile’s position and valid image area. It also records overlaps with neighboring tiles. Keep this output alongside tiles when a later step needs to understand where tiles came from or distinguish source pixels from edge-filled pixels.

bbox

A list of BBOX values giving each tile’s covered rectangle in its source image as (x, y, width, height). The coordinates identify the top-left corner, and the dimensions describe the source area actually covered, which can be smaller than the full tile at an image edge. Entries follow the same source-image and tile order as tiles.

bounding_box

A list of BOUNDING_BOX values containing the same source-image rectangles as bbox, expressed with named x, y, width, and height fields. Use this output when a downstream node expects the named bounding-box type. Its entries correspond in order to tiles and bbox.

Image Tile Batch Usage Tips:

  • Start with fixed_tile when a downstream process needs a particular tile size. Increase overlap_x or overlap_y when adjacent tiles need more shared context, while remembering that additional overlap can increase the number of tiles.
  • Choose bounded_grid when limiting the number of pieces matters more than keeping the requested tile size exact. Set max_columns and max_rows to your desired limits, then check whether the resulting, potentially larger tiles suit the next node.
  • Choose uniform_grid when you want to specify a minimum row or column count and minimum overlap. Adjust min_columns, min_rows, min_overlap_x, and min_overlap_y for that mode; changing the regular overlap or maximum-count controls will not adjust its layout.
  • Keep layout associated with the corresponding tiles batch, especially if you process the tiles before using their positions. The bounding-box outputs identify covered source rectangles, while the layout also distinguishes valid image areas from edge-filled parts of tiles.
  • Estimate the amount of downstream work before using small tiles or large overlaps on a large image batch: the node produces a tile for every grid position of every source image.

Image Tile Batch Common Errors and Solutions:

Image Tile Batch expects rank-4 IMAGE [B,H,W,C], got shape ...

  • Explanation: The supplied image does not have the four dimensions expected for an IMAGE batch: batch, height, width, and channels.
  • Solution: Connect a standard IMAGE output. If another node has changed the image’s shape, convert it back to a valid IMAGE batch before tiling.

Image Tile Batch requires nonempty dimensions, got shape ...

  • Explanation: At least one image dimension, including the batch size, is zero or otherwise nonpositive.
  • Solution: Check the upstream image node and ensure it produces at least one image with nonzero width, height, and channels.

Unsupported tiling mode ...

  • Explanation: The mode value is not one of the three supported options.
  • Solution: Select fixed_tile, bounded_grid, or uniform_grid from the mode control.

tile_width and tile_height must be positive.

  • Explanation: A tile dimension is zero or negative, so the node cannot create tiles.
  • Solution: Set both tile_width and tile_height to at least 1. The normal node controls enforce this minimum, but a value supplied through another mechanism may still need checking.

max_columns and max_rows must be positive.

  • Explanation: In fixed_tile or bounded_grid mode, one of these count settings is zero or negative. The layout builder validates them in both modes, even though they only limit the grid in bounded_grid.
  • Solution: Set both values to at least 1.

Requested overlap validation error

  • Explanation: In fixed_tile or bounded_grid mode, the requested horizontal or vertical overlap is not valid relative to the applicable tile dimension. In bounded_grid, validation also applies after the node calculates its resulting tile dimensions. The exact error wording comes from an overlap-validation helper not included in the supplied source.
  • Solution: Reduce overlap_x or overlap_y, or increase the corresponding tile dimension. If using bounded_grid, also try a less restrictive row or column limit.

Uniform-grid layout error

  • Explanation: The supplied source shows that uniform_grid calculates positions from tile size, minimum overlap, and minimum row or column counts, but its calculation helper is not included. Some combinations may therefore fail its validation; an exact error message cannot be confirmed from the supplied source.
  • Solution: Try a smaller minimum overlap or minimum count, or adjust the corresponding tile dimension. Change one setting at a time to identify the incompatible combination.

Resolved grid count does not match its tile positions.

  • Explanation: The calculated number of tile positions does not match the calculated row or column count. This is an internal layout-consistency check rather than a routine image-format error.
  • Solution: Recheck extreme tile, overlap, and grid-count settings and try simpler values. If the error persists with ordinary settings, report the settings and source image dimensions to the node maintainer.

Input image shape does not match TILE_LAYOUT source dimensions.

  • Explanation: The image passed to the tiling operation does not match the source dimensions stored in its layout. This check is present in the supplied tiling function, although the node normally builds the layout directly from the same image before tiling it.
  • Solution: Ensure that an image and its layout refer to the same unmodified image batch. If this appears while using the node on its own, report the image shape and settings to the node maintainer.

Image Tile Batch Related Nodes

Go back to the extension to check out more related nodes.
ComfyUI-Utility-Suite
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.

Image Tile Batch