I developed AI logo generator which generates images to a prompt.

When a user enters a prompt, it's wrapped into another prompt:

<lora:new15:1> newlogo {prompt}

Here is an example:
A user enters a prompt "deer", and it transforms into a final prompt "<lora:new15:1> newlogo deer".

But this approach has 2 drawbacks:
1. If a user enters something complicated, the AI can't figure out what it is. Here are some examples of that complicated things requested by a user:
- Shopify (SomeComplexBrandName)
- Hair salon (AI does not know what to generate – hair? salon? what is salon? etc)

2. If a user enters something like "a minimal logotype of a shark", it gets injected into a final prompt which in this case would look like this: "<lora:new15:1> newlogo a minimal logotype of a shark"

So I need a user to enter a subject which can be easily understood by my AI logo generator and it has to be just a subject – meaning no redundant words

If i give you some complex user's prompt, would you be able to transform it into something which can be recognized by the AI logo generator and at the same time give a user a satisfying result?

Bad example would be:
i give you: "a logo of a hair salon"
you give me: "wooden box"

i give you: "think of a logo for a platform that provides services to people who are engaged in retro drops in cryptocurrency projects"
you give me: "cryptocurrency retro drops"

i give you: "HCA"
you give me: "HCA"

i give you: "blackbird"
you give me: "blackbird"

---

Better example would be:
i give you: "a logo of a hair salon"
you give me: "hair"

i give you: "think of a logo for a platform that provides services to people who are engaged in retro drops in cryptocurrency projects"
you give me: "coin"

i give you: "HCA"
you give me: "abstract"

i give you: "blackbird"
you give me: "bird"