Argil: The ChatGPT plugin for Image Generation

TL;DR

  • Argil has launched on the ChatGPT plugin store.
  • The launch serves strategic interests and boosts visibility.
  • Argil aims to be the top player in no-code AI automation.
  • Challenges faced in the store, notably around image quality.
  • Plans to develop a multimodal chat interface.
  • Belief in the future value of ChatGPT plugins.
  • Argil is accessible for everyone to build AI automation.

You may don’t know it yet, but Argil is finally live on the chatGPT plugin store. You can now generate images directly from there, without having to learn how to prompt or connect to a discord channel.

But that’s not it, let me share with you what you can do with Argil and what’s coming your way with our multimodal vision.

Why we launched Argil on the ChatGPT plugin store?

That’s a strategic decision, the hype over chatGPT is not over yet, especially with the 2 beta features they came with the recent weeks:

1/ Plugin store

2/ Browse the web

People understand now that ChatGPT is not just a research company but also a product one with a deep interest in enabling the company to leverage the power of LLMs.

Our hypothesis was simple:

People will use the plugins, either because they answer a specific need or out of curiosity. Both these reasons are perfect for us as they provide free traffic from people falling under our ideal client hypotheses.

Our goal with Argil is to become the first horizontal player in no-code ai automation.

ChatGPT plays the role of the bridge between the mass users and other ai tools, this move got us thousands of subscribers to the platform and great insights on what steps we should take in regards to the chatGPT plugin.

The Argil version on the ChatGPT plugin store is not the best:

Unfortunately, our request was refused three times before it was accepted and one of the reasons was our use of Stable Diffusion instead of Dalle as an LLM for image generation.

We were finally accepted in the plugin store but not without a sacrifice. The quality of our images took a hit because the LLM model used on the chatGPT plugin version is based on DALL-E, a model we did not fine-tune and improve to get the best image quality possible.

This is very unfortunate as the experience propose on the plugin is far from what you could get on Argil’s studio.

Here’s one example to illustrate the quality difference between the studio in which you’ll get access to a fine-tuned and super-optimized version of Stable Diffusion models and what you can do directly on the chatGPT plugin store.

IMAGE COMPARATIVE

But are we going to stop at this chatGPT plugin? Is argil going after a more complex and complete market?

The answer is yes, here’s what you need to know:

Where are we heading with Argil’s chatGPT plugin?

As said earlier, launching on chatGPT plugin store was a way to test if virality and traction would rise from an interface that generate in natural language Images.

We did it to test our hypotheses and we confirmed it. People are familiar with the chat interface of chatGPT and have already adopted it.

This shows how powerful can it be to re-use what worked for others and personalize it to your use-case.

Our goal is to simplify in no-code the usage of generative ai: create the market of ai automation. To do that with need to get as many insights and feedback as possible and now it became very clear for us.

The workflow vision we have can only come to life if it’s done through a chat interface, at least for now. This is why we’ll push things further, and we’ll introduce shortly, the first multimodal chat interface.

Imagine, just like chatGPT but you can:

  • Call models of your trained datasets
  • Generate images with specific styles
  • Generate text out of contextual datasets

The only thing you’ll have to do is set your workflows or choose a template out of those we’ll share with you on the platform.

That’s the final form of creation: multimodal and ai native.

Is there a future for ChatGPT plugins?

As mentioned in a first article writing on this topic, we believe that chatGPT plugin store can become the apple store of ai applications. Not being present on it would decrease credibility and potential of virality.

Content is the best way to put the right eyeballs on your product, and to get these eyeballs you either create the content yourself or find a way to put your product where there’s traction.

ChatGPT plugin store is the go to for that, but will it still be?

Multimodal ai generation means out of written text you can:

  • Generate text
  • Generate imgaes
  • Generate musics
  • Generate videos
  • Generate 3D assets
  • Generate full websites

The list can go forever, even if some of these use-cases are still mature they will in the next few years become a community accessible by most.

Now what if one application centralize all of this? What’s the goal of a store that shares copy-cats with less output quality?

To reach that point, we may need more than a few years which is why I am quite confident in the presence of chatGPT plugin store in a few years and it’s ability to provide value to product builders but also to people looking for products.

Now if you haven’t tries Argil yet, what ar you waiting for? Build your first ai automation here.

Othmane Khadri