Up next in 10
AI Image Generation for Beginners
Sep 27, 2025
Getting the best Images from AI is very easy. Here is how you can do it. Create AI Images and Videos - https://lanhive.com
View Video Transcript
0:00
Getting started with AI image generation
0:02
can be overwhelming. With many tools to
0:05
choose from, it may get hard to find
0:07
which one suits your needs. But in this
0:09
video, I'll simplify the process of AI
0:12
image generation so that you can start
0:14
generating stunning images right away.
0:17
And I also give you a very simple
0:18
framework that you can use to get the
0:20
best results. Let's get started. For AI
0:23
image generation, there are only two
0:25
things that you need to care about.
0:27
Number one is the AI model and number
0:30
two is the prompt. The AI model is what
0:32
you're going to use to generate the
0:34
image and there are a lot of these. Here
0:36
are some of them. You have Imagen,
0:39
Cream, FRA, GPT image, Nano Banana and
0:44
many many others. All these AI models
0:47
are owned by different companies and you
0:49
need different platforms to access them.
0:51
But in this video, I'm going to use
0:53
LANHive throughout because it has all
0:55
the models for images and videos. So,
0:58
which one of these AI models should you
1:00
choose? Well, it's very simple. For most
1:02
of the task, imagine Cream and Fracks
1:05
are going to be enough. But when you go
1:07
into the specifics like you want amazing
1:10
text, you want text in different styles,
1:12
you find ideog
1:14
to be the best. And if you want great
1:16
images of people, you find Minax image,
1:19
Sid Dream, and Mijan are some of the
1:21
best models for this. And then in your
1:24
spare time, you can try out some of
1:25
these models because they also produce
1:27
amazing results. But for a start, those
1:30
are the ones that you should focus on.
1:32
So that's about the model. Number two is
1:35
the prompt. So this is the instruction
1:37
you're going to give to the AI on how
1:39
you want your image to be. Now, there
1:41
are many prompts out there. There are
1:42
some websites that are just dedicated to
1:44
sharing prompts, but I'm going to give
1:46
you a very simple framework that you can
1:48
use to get the best results. It's called
1:51
the C framework. So, it's subject,
1:54
environment, and action or subject,
1:57
action, environment. Now, what does this
2:00
mean? So, every time you write a prompt,
2:02
you have to make sure that you describe
2:03
your subject in details. You describe
2:06
the environment they are in and also
2:08
what action they are taking. So, let's
2:10
look at an example. So to do this, I'll
2:12
just go to images here. Then I'll do
2:15
compare mode so that we can do two
2:17
models at once and see what we get. So
2:20
I'll use Cream 4 and imagine 4. And
2:22
here's the prompt that we're going to
2:24
use. So when you look at this prompt, we
2:26
have our subject here, a black American
2:29
young lady with long hair wearing a
2:31
white crop top and blue jeans. Then we
2:34
have the action she's taking. She's
2:36
leaning against the wall. and we had the
2:38
environment on the street corner with
2:40
modern minimalistic graffiti on the
2:42
wall. So that's the safe framework in
2:44
action. So once we are happy with all
2:46
the settings including the aspect ratio,
2:49
we can just click generate here.
2:52
So we have our results here and we have
2:55
everything that we ask for in the prompt
2:57
and you can see that the two images are
2:59
similar because the prompt provided the
3:01
subject, the environment and also the
3:04
action they are doing. In both cases,
3:06
they're leaning against the wall.
3:07
There's graffiti on the wall. They have
3:09
blue hair, a white top, and ripped
3:12
jeans. So, that's one of the best method
3:14
for prompting. So, now let's try another
3:16
one with a vague and less detailed
3:19
prompt. I'll use the same mod and
3:21
imagine four, but this time the prompt
3:24
will be a beautiful lady eating ice
3:26
cream. So, when you look at this prompt,
3:28
we have the subject, the beautiful lady.
3:31
also have the action. She's eating ice
3:33
cream, but we don't have the
3:34
environment. So, the AI would generate
3:37
its own environment. And there are also
3:39
many other details that are missing
3:41
here. We don't know whether the lady is
3:42
standing or sitting or walking. We also
3:45
don't know what she's wearing and many
3:47
other things. So, we have left all those
3:49
to AI. Meaning the AI model is at
3:53
generate whatever it deems right in such
3:55
a case. So in such a case you realize
3:58
that the tool modules would generate
3:59
images that may not be similar like the
4:02
previous ones even though the prompt is
4:05
the same. So we have no control of many
4:07
things here. But let's hit generate to
4:09
see what will happen.
4:12
Okay, we have our results. So you can
4:14
see that the tomb images are widely
4:16
different. The environment is so
4:18
different because the environment was
4:20
not specified. The clothing is also
4:22
different and many other things because
4:25
we left some of the details for the AI
4:27
to figure out. So this is not the best
4:29
way to prompt. The best way is to
4:31
describe your environment properly, the
4:33
action your subject is taking and also
4:35
the subject. And if you want to try this
4:38
platform, I'll leave the link in the
4:39
description below where you can get some
4:41
free credits to try it. I'll see you in
4:44
the next one.