Dreamlike AI Review: Is AI Art Worth the Hype?

AI generated art is killing creativity! You may have heard people voicing this opinion, and it’s fair enough to be worried about the advancements in artificial intelligence. However, AI art generators can actually improve and encourage creativity rather than hinder it. 

With apps like Dreamlike Art, you can visualize anything you want. And we mean anything. Writing a book and want to see the characters? Or maybe you’re an artist wanting inspiration. Whatever your reasons, the power of AI can be the way forward.

So, we’re testing out Dreamlike Art and seeing whether it’s a tool worth bragging about. We’ll assess the pros and cons of its features and share the original images we created, so you can decide whether you should invest. 

What is Dreamlike Art?

Dreamlike Art is a text-to-image AI website. Using natural language processing (NLP), the AI can understand prompts written in conversational English and generate digital art based on the instructions. 

Within the prompts, you can specify what kind of artistic styles you want to see, from acrylic painting to fine art. They are high-quality images, meaning if you want art prints on t-shirts, canvas prints, or simply to share on social media, then all you have to do is download them.

Trying it out

We’re going to take you through our experience step-by-step of using this app and the results. We had no prior knowledge of its features, so you can get an idea of just how user-friendly it is. 

Getting started

So, initially, we were a little confused. There are no tutorials or explanations to shed light on the many options you see on the main dashboard. Granted, we were able to start for free without a credit card which was a bonus. All we needed to do was create an account which took 5 seconds.  

Dreamlike AI's dashboard

Generating an image

We thought we’d learn by trial and error. Initially, we just filled in the ‘prompt’ box and left everything else blank. We asked it to generate a “painting of a very happy cat eating spaghetti in natural lighting with a soft color palette.” Below are the results.

Two ai generated images of a cat eating spaghetti but the fur and spaghetti are mixed and it's quite frightening.

While slightly terrifying, the color and the lighting were spot on and they were visually captivating images. However, it seemed to have trouble differentiating the spaghetti and the fur and knowing where the legs and paws should be. 

Trying the ‘edit’ feature

We decided to try out the ‘edit’ function to see if I could correct these issues. So, I wrote in the edit box: “Each leg and paw should be on either side of the bowl holding a fork. The fur shouldn’t have any spaghetti in it.” I thought I would try out the ‘negative prompt’ box, which I assume means you enter what you don’t want included. Here, I wrote: “deformed cat, bad anatomy, and spaghetti fur.” 

Unfortunately, this feature didn’t work. It generated two of the exact same image as I started with. So, instead, I went back to the drawing board and entered more detail in my initial prompt, and included a negative instruction. 

The edit feature with positive prompts and negative prompts added.

This was a lot more effective. There were fewer floating paws and no spaghetti fur. Result! 

Two ai generated images of cats eating spaghetti without any fur and spaghetti mixing.

The images generated weren’t perfect, but it was definitely an improvement. I thought I’d try out the edit feature once more to see if I could get the image on the left to include one spoon. Bad idea. See the results below. 

AI generated image of a cat made of spoons.

Trying the ‘initial image’ feature

Next, I tried out the ‘initial image’ feature. Here you can upload an image, and the generator will draw upon it for inspiration when creating your image from the prompt. Below you can see I uploaded a previously generated piece of art and wrote in the prompt: “Can you add a dog eating pizza next to the cat in the initial image.” 

Trying the initial image feature with an example prompt and negative prompt and an image of a cat eating spaghetti.

Again, the result wasn’t great:

Two ai generated images of two cats eating pizza but they are turning into the pizza

Trying the different AI models

Dreamlike Art offers eight different models that will generate different types of images:

  • Kadinsky 2.1
  • Neurogen 1.0
  • Realism Engine 1.0
  • Counterfeit 2.5
  • Dreamlike Diffusion 1.0
  • Dreamlike Anime 1.0
  • Dreamlike Photoreal 2.0
  • Stable Diffusion 1.5

Previously, we had just used the default, Kadinsky 2.1. So, we decided to enter the exact same prompt and negative prompt, trying out a few of the different models. 

Realism Engine 1.0: 

Two ai generated images of cat's eating spaghetti in a realism style.

Neurogen 1.0:

Two ai generated images of cats eating spaghettii in the neurogen style.

Dreamlike Diffusion 1.0

Two ai generated images of cats eating spaghetti in the stable diffusion style.

I think this model was the most successful. There are no errors like floating paws, and the spaghetti is well-defined. It’s very stylistic, but it’s also high quality. 

The model you choose will depend on the type of image you want to generate. For example, the Counterfeit 2.5 and Dreamlike Anime 2.0 models seem to specialize in generating Anime characters. Whereas the Dreamlike Photoreal 2.0 model generates photographs. 

Pricing

So, you’ve seen what it can do, but how much will it cost you? The free sign-up awards you 50 credits to splash. Generating two images costs 4.4 credits, so you can have a good play around and get to grips with the software before making any decisions about money. 

After the first day of the free sign-up, you can get 24 credits a day. We think this is pretty generous, although it’s worth noting that every edit also takes up credits as if they were new images. And, as you may have realized, it may take a good few tries to generate the image you want. 

If you need more credits, they have three different paid plans. The cheapest starts at $12 a month and allows you to use 3000 credits a month. The next package jumps up to $24 a month, giving you a whopping 12,000 credits. After this, the Ultra Plan offers you 30,000 credits at $48 a month. Importantly, these prices I have mentioned have a 20% discount already applied to them on the assumption you will be billed yearly, but there is a monthly option. Each package includes access to different features, so check out the full details below. 

4 different pricing plans: free plan at $0 a month, Core plan at $12 a month, pro plan at $24 a month and ultra plan at $48 a month.

What do we think?

It’s clear from our attempts that Dreamlike Art may not be the best go-to app for more professional projects. However, it’s still a fun website to play around with and can create some stunning original art with some practice. Once you learn what prompts work best, the results will likely be more successful. 

The AI software will only become more advanced, so it’s a good idea to keep an eye on this app. It’s also worth noting that once you get the hang of the features and their capabilities, you may find a way to make this website work for you. 

There are plenty of other image generation out there to try, so consider checking out the following:

This is certainly going on our list of powerful AI tools. If you want a full list of AI tools, head over to our extensive library to learn more. In the meantime, sign up for the Top Apps newsletter to stay up to date with this ever-advancing technology.