Check out my prompting guide to get much better results then a LoRA will ever be able to do:
Controlling "Zoom" With Words Alone

Krea2:

This is much better compared to the Z-image version. But it is extremely dependent on your prompt. It will NOT override framing if you have that in your prompt. But if the prompt leave framing and related details out, this Slider will do some extreme zoom.
This is NOT a wide angle slider, this is pure zoom. It will enhance fisheye shots etc, but it will not just make everyone distorted like a wideangle lens would.
It does mess a little with colors and quality, but obviously this should only be used on the first sampler in your workflow.

Recommended range stable: -5.0 to 5.0
Absolute limits (unsure): -12.0 to 12.0
Range will change with different prompts.

ZIT:
Closeup / Wide Shot HELPER Slider

This will not zoom in or out your image no matter what. But it will help you to get even wider shots, or more closeup when combined with prompting for it.
For example, if the prompt say "a wide shot" this slider will not just override that and make it a cluseup, but it will make the wide shot less wide.
You can see from the samples that the effect is neither stable nor perfect. But from time to time i feel like its impossible to get Z-image to zoom out just a litle more, and that is what this LoRA is perfect for.
The line between total collapse and perfect effect is usually very narrow, so one value can work perfectly on one image, and fall appart on another prompt.
This is by far the hardest LoRA i have made, and i am in no way happy with the result. But this is the best i managed to do. The concept is mainly closeup/full body, so the effect is limited to that range in general.

This is the type of LoRA you dont use on a day to day basis, but when you first need it, its very usefull.

Recommended range closeup: -6 to -7 (+1/-1)
Recommended range wide shot: 6 to 7 (+1/-1)