The recent hype over AI made it all over the place, with chatGPT and Midjourney being the two top contenders. Respectively the first is for text generation and the second for image generation.
To be fair, I truly believe not a single person in the tech ecosystem hasn’t heard of both of these tools.
The reason for that is simple, Midjourney is able to render top-quality results of image generation in comparison to other competitors and chatGPT is pushing the limitations of common chatbots to a whole another level.
Today we’ll focus on a specific frustration experienced by product builders, marketers in big companies, ai consultants, creative agencies, and many other Persona:
The official Midjourney API is not yet out.
That’s a shame when you see the quality of the outputs rendered by Midjourney and the horrible user experience you need to get through for image generation….
Midjourney positioning is pretty clear, the fact that they didn’t create Midjourney API can only mean one thing:
They focus completely on building the most advanced model of image generation. They want to destroy their competitors and be the only survive.
Pretty egotistic if you ask me, playing the game of the horrible UI and seeing if it still attracts people to their product instead of competitors is a good definition of self-centrism.
Midjourney is not a product company, it’s a research and development one, having Midjourney API available for everyone may not be in their plans now, but it’s a risky move as competitors did not say their final say
Currently, any person that wants to leverage Midjourney to their specific use case will have to create a discord account and generate the image in a specific chat.
The user experience is daunting, and not scalable. While mid-journey nailed the research components to build the most advanced image generation foundation model, they’re missing the product part of their business.
Some smart builders saw an opportunity here and build an in-between API to use bots. These bots will interact with the Midjourney Discord chat for you and you’ll directly get your image from another interface.
Still, that solution is not scalable and the experience is not welcoming from all types of profiles, which is exactly why Midjourney API is that awaited by people.
Will they do it?
I don’t know to be honest, and at Argil we decided to not care. Midjourney is not the only company building a foundation model.
1 month ago we thought that the race for the best model was already won by Midjourney, but Stable Diffusion just released a week ago a new version of their model, and well….
The results are astonishing, we’re near Midjourney's realistic quality.
Here are a few examples:
Well, unfortunately, we’re not using Midjourney directly but with the latest update of SDXL, the results provided by the model are very near to what Midjourney can do.
And for that little difference in quality, we have a lot of features to convert your doubts into excitement to test our Image studio. At Argil we have one straightforward goal in mind:
Make AI usage easy for all by building a horizontal application for ai automations.
Here are a few features you won’t have with Midjourney or the application that promises they provide you with Midjourney API:
1/ Train a model on your datasets
With Argil you can upload images of yourself, for example, to create a model that knows how you look like. The ai will learn from the images you put as inputs and you can then generate Midjourney-like results but with yourself as the model of the generation.
2/ You can create templates
Let’s say you’re going after a specific industry: Retail boots image generation.
You may have generated Thousands of different images, and 50 of them are the best. You can select these images in Argil’s studio and build a template. You can then use the template when generating new pictures and the studio will take it as a base. This will increase the quality of the image you generate.
3/ Full ai-automation
On Argil, you have the ability to build ai-automated workflows that can then be streamlined to your application via our API. This means that any workflow using image generation that you want to build with top quality that so far required human intervention can be automated on Argil.
Midjourney API might not be officially out, and other foundation models might not yet provide 100% of the same quality as Midjourney, but try Argil and judge by yourself.
The use case enabled far outweighs the small existing differences in image quality, try it here.