As far as AI image generators go, it seems a new application is launching every day. Whether it be discord based, app based or running off its own domain, there are plenty of programs to choose from. But what I’ve come to find is many of these programs are free to try, but soon require a paid membership to continue generating.
That was until I discovered Playground AI, a domain based image creator that has exceeded my expectations. It has many of the same features of MidJourney, but with no attached cost.
With unlimited generations, it’s perfect for someone looking to experiment or produce large amounts of content. I find the first dozen or so images generate in about five seconds, and although it will eventually display a banner claiming that your generations will take longer on the free plan, I have never had an image take longer than thirty seconds.
There are 38 filters to choose from if you’re trying to acquire a certain style, but selecting ‘none’ and simply putting in the description what sort of style you’re looking for seems to work just as well, if not better.
You also have the option to put negative keywords, things you don’t want to appear in your image. Lets say the AI keeps generating a motorcycle in your images, even though it has nothing to do with what you want. You can add the word to the section, and it will no longer appear.
It also has an extremely impressive image to image tool, where you can either select an already generated image, or upload your own, and choose from a sliding scale of 1-100 on how similar you want the about-to-be-created image to be.
Now you may be wondering how to get an image generated on Playground to pop up in Image to Image rather than an uploaded file. Simply hover over your desired image and select image to image. You will also see several more options that are crazy to see in a free version, including background removal, variant creator and upscale.
Playground AI runs off of stable diffusion, which is what helps it remain free. You can choose from either Stable Diffusion XL or Stable Diffusion 1.5. XL is far superior, but 1.5 is fun to experiment with anyways.
Playground also allows you to tweak your images size without having to add it to your description. There are six dimensions to choose from, including
- 512 x 512
- 1024 x 1024
- 640 x 384
- 384 x 640
- 768 x 512
- 512 x 768
Now onto more cool features I am a huge fan of. There are three more sliding scales you can use, although you can only use so much at once. Since you are on a free version, you can’t be using every feature at once. So in order to use some of these sliding scales, you’ll need to be only generating one image at once, or have nothing else fancy going on like image to image.
One of these scales is Prompt Guidance. From 1-30, you can choose how closely you want the AI to follow your prompt. The program recommends 7-10, as this still gives the AI some freedom, but the higher you set it the more it will take your words to heart.
The other scale is Quality and Detail, which runs from 10-150, which can drastically improve your image, but can take much longer to generate.
Much like MidJourney, Playground has a seed option, which allows you to assign a number to your image, which can be used to create variations of your image or play off of it, similar to the image to image feature.
On the free version, you can generate 1-4 images at once. The more images you select, the longer it takes, though it does cut down on the need to keep hitting generate to try and achieve the perfect image.
I find Playground is also excellent at generating PNG capable images. Although it does have the ability to remove the background, it sometimes cuts edges or doesn’t produce a perfect cut out. Though it can provide an image that is easy to put into a photo editor such as Canva or Fotor, and remove the background. It also doesn’t generate a complicated background on many of the generations if it does (usually just a solid colour) that include the word ‘PNG”, which I find can be a struggle for other platforms such as Leonardo.
This can be great for generating logos or logo components, or you can use it to generate a complete logo all together. Although you won’t be able to achieve proper text integration just yet. AI still seems to produce gibberish combinations of letters, but it’s getting better.
Though you can try to get a logo generation without text, you simply just have to add “no text” to the ‘Exclude from Image’ section. Though this setting isn’t fool proof, and may still throw in some words, it will at least provide a generation or two featuring no text.
Now, onto some of the bigger downsides of the free version. By not having the Pro Plan, your images aren’t private, per say. Although it isn’t like some AI generators that post your generations to a public chatroom. Instead, they may appear in the ‘Community Feed’, although it is unlikely to see your image unless you are specifically searching for it. Though for someone who wants complete privacy, you can subscribe for a reasonable price.
I myself am still experimenting with and learning what Playground has to offer, and I’m sure new features are set to drop. But for now, consider Playground.AI a strong free contender for AI image generation. Happy creating.
