Welcome to my 💫🌄 Friendly Ideogram 4
✨ Less mess, more magic and control
Feel like a real designer with Ideogram 4 and my friendly workflow!
I offer my workflow with LLM Enhancer & Prompt Builder modes. Compose each part of your generation!
The latest version of ComfyUI is required.
💻 System requirements:
Minimum system requirements:
RTX 3000-s, 8GB+ VRAM, 32GB+ RAM, 8-core processor, SSD, latest ComfyUI
📌 Detailed tips and links to models in the workflow
✨ Workflow features:
Extremely user-friendly interface
Maximum performance even on low-end systems with at least 8GB of VRAM
LLM Enhancer and Prompt Builder implementation
Support for up to 2 Lora
Detailed tooltips and links to models
Manual random seed for complete control over generations
Model overview
Ideogram 4.0 is a powerful, open-weight text-to-image model designed for professional-grade visual creation. It features a 9.3 billion parameter single-stream Diffusion Transformer (DiT) architecture that was trained entirely from scratch
Key Capabilities & Features
Production-Grade Typography: It offers crystal-clear, dense text rendering across multiple languages directly inside the composition, rather than as a fragile overlay
Native Transparency: The model can generate images with native transparent backgrounds, eliminating the need for post-processing background removal
High Resolution & Flexibility: It supports native 2K resolution output and flexible aspect ratios up to 2048 pixels on each side
Advanced Controllability: The model is trained on structured JSON captions, providing unprecedented control over composition, style, lighting, and color palettes. It also allows for precise object and text placement using bounding box controls
.Post-Generation Editing Suite
Ideogram 4.0 introduces a comprehensive suite of tools to revise and refine images after generation without starting from scratch:
Layerize Text: Turns generated typography into editable layers so you can restyle, reposition, and refine the text afterward.
Prompt Edit: Allows users to describe changes in plain language to revise specific parts of the image.
Extend & Reframe: Grows the composition beyond its original crop or shifts it for different aspect ratios while maintaining the original focal point.
Magic Fill: Adds new objects or replaces missing areas to seamlessly match the surrounding scene.
Other Tools: The platform also includes native background removal, upscaling for sharper delivery, and remixing to explore new creative directions.
🤗🙏🏼 Thanks to Ideogram AI and Kijai
Original repo — GitHub
Description
27.06 (v1.2.1) change log: more LLM optimizations and some fixes
2x acceleration with new SageAttention (links in the workflow)
Optimized LLM for complex scenarios (for example for extreme angles)
Added random seed for LLM
Updated links and tips
FAQ
Comments (3)
I assume in this workflow, the prompt box has to be empty for the image reference to be used? Also, I noticed that no matter what sizes are set, the preview in the builder doesn't change to match.
@ByteCrafter if you need maximum similarity to the reference image you can leave prompt empty or write something like "create exactly the same image, keeping all the details". If you want for example to change background or character pose you can write "make this man running on a beach". As for the preview in Prompt builder editor it must change according to aspect ratio in my workflow with default settings. The logic is this: you write your simple prompt with LLM Enhancer turned on, set resolution and aspect ratio in Easy resolution node, then run generation. After LLM generates json prompt it enters in Prompt builder automatically and reads the prompt and resolution. You should make sure that Always mode is activated in Prompt builder first. Only in this mode Prompt bulder can get json prompt from LLM automatically. Also I reccomend gemma e4b fp8 model because of stable result. Other lighter models can generate incorrect json prompt and Prompt builder can't work with it.
So and if you get json prompt in Prompt builder you should change mode Always to When empty. After that you can change everything in Ideogram 4 Editor window and also resolution and it will be reflect on next generations
@RusselX I did notice that no matter the prompt style, the JSON always sets the style to photo; it never chooses art_style. I had to create a new system prompt on a switch to switch between photo and art_style. - Cheers for the heads up on the image as a reference system.