API Guide

Ideogram API Guide for Ideogram 4

A focused non-official overview of how Ideogram 4 API workflows relate to text prompts, structured json_prompt inputs, and describe-v4 image-to-prompt output.

What this page covers

Focused topics, kept narrow so the page stays useful.

ideogram apiideogram 4 api

Use text_prompt for fast ideas

A text prompt is useful when you want a quick generation without planning every element. For simple portraits, simple illustrations, or broad creative exploration, this is enough.

Best for fast exploration.
Lower setup time.
Less precise for typography and dense layout.

Use json_prompt for planned layouts

A structured json_prompt is better when the image needs exact text, repeatable style, color palette control, or element placement. It behaves more like a design brief than a casual prompt.

Separate text from visual description.
Use bbox when placement matters.
Keep JSON field names and order close to official guidance.

Use describe-v4 to convert images into structure

The official describe-v4 endpoint can return a structured JSON prompt from an uploaded image. This is useful for learning how a composition can be represented before generating a new image.

It can preserve bounding boxes by default.
The returned json_prompt can be reused in v4 generation endpoints.
Always verify official pricing and terms before production use.