What are the best online tools for AI-enabled image creation or editing that have more interactivity?

So, not just "you write text, wait 30 seconds, image comes out", but more two-way feedback between user and diffusion algo, ideally without 30 second latency at each step.

Unfortunately online tools are quite limited in that regard.

Your first, and best option is installing AUTOMATIC1111 and then customize it with plugins: https://github.com/AUTOMATIC1111/stable-diffusion-webui (but it takes a lot of time)

Thread..

3,
Among online tools, DALL-E is a pretty good one, you could talk to ChatGPT and it will help you generate exactly the images you wanted: https://chat.openai.com

4,
One of my personal favorite if the Nouns AI Playground: https://www.nounsai.wtf/ it works brilliant for "controlnet" to create different versions of existing images.

5,
And finally Midjourney is still a good option, but It's not a very advanced tool in terms of plugins and extensions. It's pretty much WYSIWYG.

6,
I'd recommend you to ask someone to setup and self-host a private Automatic1111 or ComfyUI instance for you, so you could control how fast it is and what it does exactly. The plugins are capable of generating videos, advanced controlnet-type manipulations, LoRa (extension models) etc. All the professionals do this.

+ Extra: If you get deep into the rabbithole of image diffusion, you'll want to train your own LoRa models, and the best tool to do that is a self-hosted "kohya-ss": https://github.com/Akegarasu/lora-scripts

Powerful, and it took a week to understand what I'm doing, but only 1 hour to finish a model after dataset.

2,
ComfyUI is the second best option for self-hosting: https://github.com/comfyanonymous/ComfyUI it has the most creative web interface.