How to use Ai to update an Ai image?

Questions
1

I am using the OpenAi Dall-E to create an image, however I notice that some things may not be exactly as I want them on first try, such as text might be missing a letter or background needs adjusting.

I like the first version that is created, but just want to tweat/adjust the image.

What are some recommendations for getting the original image into the prompt for it to be able to output something very similar just with adjustments?

Ive tried using the ai image analyzer and then feeding that in as a prompt, but the image comes out different and with a whole set of new issues.

n8n version: 1.53.2
Database (default: SQLite): dont know
n8n EXECUTIONS_PROCESS setting (default: own, main): dont know
Running n8n via (Docker, npm, n8n cloud, desktop app): docker
Operating system: mac

2

It looks like your topic is missing some important information. Could you provide the following if applicable.

  • n8n version:
  • Database (default: SQLite):
  • n8n EXECUTIONS_PROCESS setting (default: own, main):
  • Running n8n via (Docker, npm, n8n cloud, desktop app):
  • Operating system:
3

This question is beyond using n8n and very specific to how LLMs work in general. You ran into a big issue with AI generated content - output is generally ‘unstable’ and minor prompt changes can have unexpected impact. The reverse analysis that you’re trying will make this even worse by adding more randomness. Your best bet is tweaking the original prompt, but even that will probably not do exactly what you want.