comfyui-workflows/models/ideogram/README.md

Ideogram Workflows

Use this folder for Ideogram workflows, especially anything focused on reliable in-image text, poster layouts, labels, flyers, or other typography-heavy compositions.

Strong current lead

• Ideogram v4: Text to Image (API)
  • Source workflow page: https://comfy.org/workflows/5ae006b13a36-5ae006b13a36/
  • Model listing page: https://comfy.org/workflows/model/ideogram/
  • Partner-node docs reference: https://github.com/niknah/comfyui-docs/blob/main/tutorials/partner-nodes/ideogram/ideogram-v4.mdx

Why it matters

• The workflow is explicitly built around structured JSON prompts instead of only free-form text.
• The prompt format supports:
  • high-level scene description
  • style description
  • literal in-image text
  • color palette control
  • per-element bounding boxes for layout guidance
• This is the strongest current lead for text placement because it targets the exact weakness ordinary prompt-only workflows usually have: getting legible text into the right part of the frame.

Practical notes

• Ideogram 4.0 is documented as supporting either natural language or structured JSON prompts, but the structured JSON route is the one to prioritize for posters, labels, and other text-first layouts.
• The workflow notes say negative prompts are not needed for the partner-node path.
• The docs also mention an optional prompt-builder path for authoring schema-matching JSON more comfortably.

What to import later

• Prioritize the Ideogram v4: Text to Image (API) graph first.
• If it imports cleanly, treat it as the baseline text-placement workflow for comparison against Krea 2, FLUX, and FLUX.2-klein.