Runware

Ideogram Layerize Text

Extract editable text layers and a text-erased base image from design visuals

Launch model

Ideogram Layerize Text Overview

Ideogram Layerize Text analyzes an image, detects readable text regions, and turns each line into structured editable text blocks with position, styling, and font information. It also returns a text-erased base image, making it useful for posters, social graphics, packaging, book covers, and other design workflows where copy needs to change without regenerating the whole composition.

How to Use Ideogram Layerize Text

Overview

Ideogram Layerize Text is a design-focused utility model that extracts editable text layers from an image.

Instead of leaving typography baked into a flat raster output, it detects text regions and returns structured text blocks along with a text-erased base image. This makes it useful for design iteration workflows where the layout should stay intact while the wording, font choices, or text sizing change.

Strengths

Editable Text Extraction

The model turns detected text into structured text blocks that can be selected, rewritten, resized, and restyled outside the original flat image.

Text-Erased Base Image

Layerize Text also returns a base image with the detected text removed. This makes it easier to rebuild the design with editable typography while preserving the surrounding visual composition.

Design Workflow Utility

The model is especially useful for posters, book covers, social graphics, packaging, flyers, merch graphics, and other typography-driven visuals where the design should survive copy changes.

No Regeneration Required

A typo fix, copy rewrite, font swap, or localized headline can be handled without rerunning a full image generation workflow.

Structured Styling Data

The returned text blocks include details such as text content, positioning, alignment, font information, color, formatting, and role labels, which makes the output useful in programmatic design systems and editing pipelines.

Capabilities

Image-to-Text

The model extracts text content and structured text-layer metadata from an image.

Image-to-Image

The model returns a text-erased base image that can be used to rebuild or edit the design with separate typography layers.

Input and Output

  • AIR ID: ideogram:layerize-text@0
  • Input: an image in JPEG, PNG, or WebP format
  • Output: structured text blocks plus a text-erased base image
  • Best with: clear, straight, conventional typography

Best Fit

  • Poster and flyer workflows
  • Social media design iteration
  • Packaging and label updates
  • Book cover copy changes
  • Programmatic design pipelines
  • Localization and copy testing