Six months of working on evenings and weekends finally brought my idea to life.
Countless bugs, code rewrites, varying levels of motivation. I had no idea it would take so much effort to finish what seemed like a very simple concept. I was naïve enough to think I will have MVP in 2 months, but it ended up taking three times as long.
But I finally made it. The application, though not entirely bug-free, is up and running, and quite functional.
While the success of it is uncertain, and the timing for AI app might not be the best, I consider this a small personal accomplishment.
Before diving into the details of the idea, let me give you some context about myself.
The platform I built is for creating virtual influencers using AI.
There are at least a couple of scenarios where virtual influencers can be beneficial:
There are at least a couple of scenarios where virtual influencers can be beneficial:
So, seeing the potential, I have build a whole platform for this purpose.
At first user generates a face:

And then can use it to create images with consistent face:

That can be assigned to the influencer and accessed anytime:

AI suggested content ideas, and prompts for image generation, inpaint (ex. necklace generation on a image), are on its way.
To clarify, I didn’t create the image generation software myself—I’m using Fooocus as the core technology.
While I could have used Stable Diffusion, which would have made my life easier (easier to implement and cheaper to run as a serverless function), I found Fooocus delivers better results out of the box for generating influencer images.
This what I use for the backend:
And for the front-end:
Since I have no experience with JS frameworks, I stuck to the basics.
Choosing the right hosting is a topic worth its own post—I've tried various options to find the cheapest solution. In short, I’m using a GPU VPS to run the image generation software and a separate server for the web application.
Yes, I’m aware that AI-based SaaS platforms are very common nowadays. My idea could be seen as “just another AI wrapper,” and that’s partially true—except I’m adding extra value to the image generation software.
The key difference is that I aim to make this a comprehensive platform for virtual influencer generation. This includes:
But getting back to the reason why I chose this idea, even though it might be too late to succeed in this niche:
I’m always trying different ideas because that’s how you grow—you have to start somewhere.
After my last idea (short documentary on a cybersecurity topic), I spent about two months searching for a new idea worth pursuing. That’s a long time being inactive, so I was eager to start doing something.
I accidentally found an app of one of my competitors. The idea of creating a digital influencer platform, with a community of like-minded people and tips on growing your AI influencer account on social media, seemed having potential to me.
So, I started working on this idea at the end of January of 2024. As you can see, it took a while to get everything up and running.
Honestly, this is the most undefined and challenging aspect for me.
When shipping new features, everything depends on you—you're there to solve the problems, and with enough dedication, you will eventually solve them.
But once you’ve finished building and released your idea into the wild, everything depends on whether users like it or not.
Here are a few ideas I aim to implement:
What I decided not to - submit to AI tool directories. Listing on one of the most popular directories now costs $347, which I find insane. I don’t see much value in paying that much.
If you have any specific questions, please feel free to ask.
I would also really appreciate any feedback on the product or advice on marketing.
I plan to occasionally share updates about my marketing efforts and results on X. https://x.com/gsaulenas.
If you'd like to join a newly created community on virtual influencer generation, you can find it on Discord - https://discord.gg/JkwwVCwJ.