What We Built
TL;DR:
We integrated OpenAI’s new gpt-image-1 API (from GPT-4o) directly into a fully functional visual editor, our CreativeEditor SDK (CE.SDK).
Instead of prompting in ChatGPT and manually copying results, users can now generate, edit, and refine images inside their creative workflow - no switching between tools.
Visual Prompting
The most important bit of this demo is neatly tucked away on page 5. You can add plain text instructions and reference images on your page and compose them into a visual prompt.
This enables a true multi-modal workflow:
Feeding this page as image to the gpt-image-1 API gives us the following result:
Why Embed AI into an Editor?
Models Alone Don’t Match Real Workflows
The dominant interface for multimodal model such as gpt-image-1 is still the chat. But real-world creative tasks need more than a text box, they need placement, refinement, and iteration across different assets and formats. This is an attempt to break these models out of isolation and make them more useful to developers looking to build creative workflows.
Orchestrate Any AI in One Interface
Embedding AI generation inside CE.SDK lets you combine multiple models (whether text, image, video or audio) into one editing environment switching seamlessly between prompting, editing, and fine-tuning visuals.
From Prompt to Pixel - In Place
Users can now:
- Generate images from scratch
- Edit existing images via prompt
- Mix and combine multiple generated or uploaded images
- Use entire compositions (images + text annotations) as visual prompts for new generations
This moves AI beyond static generation toward real-time creative augmentation.
What This Unlocks
“AI needs to break out of the text box.”
By embedding image generation directly into the creative environment:
- Designers and content teams build with AI in context, rather than toggling between disconnected tools.
- Creative teams can automate localization, personalization, and variant generation at scale.
- It makes multimodal AI usable for real-world asset creation workflows not just for concept generation.
Integration & Feedback
This demo is rate-limited , if you would like to test more extensively or if you are interested in giving the AI editor a spin inside your own app, you can easily input your own api key in this demo or follow the getting started documentation here.
We’d love your feedback, any thoughts, questions, and ideas are welcome!
Reach out to us at ai@img.ly
