AI image generator Stable Diffusion already has a whole lot of fans, and now those experimenting with the brand new AI technology to develop their very own models have a spot to share their work with other enthusiasts. A startup called Civitai — a play on the word Civitas, meaning community — has created a platform where members can post their very own Stable Diffusion-based AI image models for others to find, in addition to the output of their work — AI photos — for consumers to browse and luxuriate in.

Explain Civitai CEO Justin Maier, the concept for the startup got here about because he identified there was a necessity for a spot where people could share their models and others could find them. People, he said, would post images that they had made but others didn’t know the best way to duplicate their work to make their very own images.

After wrapping up a project at Microsoft, where he had worked as a contractor on web development projects, Maier found himself intrigued by Midjourney.

“It scratched that very same itch that web development had where I saw something — or I had something in my mind — and I could type some text in get something back. It became this collaborative experience where I could explore my creativity and I get pulled in directions perhaps I hadn’t planned,” he says.

But Maier soon found himself limited by Midjourney’s credits-based plan that ultimately had him waiting for five minutes for every image to be generated. That, he says “was an actual bummer because I had fallen so deeply in love with this ability to explore and create anything that I could consider,” Maier tells TechCrunch.

Around the identical time, an open-source version got here out called Stable Diffusion that mainly means that you can do the identical thing by way of AI image generation. Because the community around Stable Diffusion grew, people were learning the best way to do various things with the model through prompting or by adding recent concepts into the model, like using images of themselves to make AI-generated selfies, for instance. Maier began to see people posting their very own models that might allow you to do cool things — like generating in a synth wave style or a punk style — and merging models together to make other recent concepts, as well. But these were being posted around the net in places like Reddit or Discord, not in a centralized community.

That led him to create Civitai (or Model Share, because it was originally called), a web-based community that organized this content in a way where people could come back and find it later.

Initially, he seeded Civitai by reaching out to model creators and asking them in the event that they could post their work on the location. At first, there have been only around 40 or 50 different models available. But over time, the location began to grow.

“People were thrilled to be seen for this effort that that they had made. And so by about January, we had turn into the go-to place for sharing these items,” says Maier. “There have been other sites that folks were posting to, like HuggingFace…nevertheless it wasn’t image-centric,” he explains. “So we became the de facto standard for sharing your model and various AI resources and the pictures that you simply’d made,” Maier adds.

By January 2023, the location hit 100,000 registered users and it occurred to Maier that it may very well be greater than a community project — it may very well be an organization. So the Civitai was then “officially” founded, and three months later, it hit one million registered users. Today, that number is around 3 million registered users and it sees around 12 to 13 million unique visitors every month.

“It’s wildly how quickly it’s grown — definitely larger than we thought it will be back in January,” Maier says.

Of the three million users, only around 10,000 unique creators every month are literally uploading recent models. Nonetheless, that number has risen by roughly 25% during the last month, after Citivai added the flexibility for people to coach recent models on the location, which makes it easier to start. A bigger variety of users are consuming the models and the content created.

Image Credits: Civitai

In consequence of its growth, Civitai, which can be co-founded by Maxfield Hulker and Briant Diehl, raised a $5.1 million round in June led by Andreessen Horowitz (a16z) at a $20 million valuation. The opposite participants within the funding were a bit unusual — the legal team that helped them close the round also got on board.

“Civitai is the prime example of an organization that’s already built an incredible, engaged community and all without spending a dime on marketing,” said Andreessen Horowitz Partner Brian Kim, in an announcement. “Our investment in the corporate will only supercharge something that’s already working incredibly well, by contributing to a world where any individual can benefit from what the largest technological shift of our generation – AI – has to supply.”

To make use of Civitai, users can upload a series of images that represent the style those images represent, then select the bottom model — like the usual model, anime model, or more realistic model, and so forth. After about an hour, the brand new model must be ready and also you’ll have the opportunity to generate your personal images using that style you had captured. The photographs generated on-site include metadata that detail things just like the prompts and resources used, and Civitai encourages those that generate models off-site to achieve this, too.

There are issues with artists checking out their work has been used to coach AI models, which is a priority. To handle this, Civitai has created a process that enables artists to flag resources they consider are using their work, which kicks off a negotiation as to the following steps. In some cases, that’s the removal of the resource entirely, or other times, the artist just wants their name removed.

“Ideally what we wish to have occur — eventually — is that these artists can…use these styles for themselves. And in the event that they want to present people the flexibility to create of their style, then they may give people the flexibility to pay to try this,” says Maier. “Nevertheless it’s still early days and I haven’t had a chance to work with many artists which might be considering doing that,” he admits.

404Media points to a different more significant issue Civitai has — it allowed users to post non-consensual porn images, and created a bounties feature for making AI images of certain targets, styles or compositions, paid for in a virtual currency. The corporate’s policies prohibit non-consensual porn AI images, but don’t limit using training AI images on real people. The report indicates that users found out the best way to piece together models to then share the non-consensual images off-site. The location’s investigation indicated that it had seen Civitai each implement the policy and ignore it, at times.

Civitai responded to the report by noting its policy prohibits mature and suggestive content, offers a removal request form and email address, and says it doesn’t allow using resources intended to breed the likeness of real individuals with its on-site generator. (The flexibility to generate real people is obtainable with the foundational Stable Diffusion model, nonetheless.) It also noted the “bounties” should not publicly available, but when the resulting images are shared on-site, they need to abide by the present “real people” policies. Bounties that ask for the likeness of real people represent 10% of requests, it said.

Despite wading into this treacherous area, Maier believes there may very well be a marketplace for legitimate AI work in the longer term, which would come with using AI imagery inside things like movies, music videos, applications, and other creative endeavors. Within the near term, though, the plan is to create a consumer-facing mobile app that may work as a repository of the AI imagery, as a form of companion to the experience happening on the major site.

The corporate can be now planning to give attention to allowing users to monetize their work, whether that’s connecting them with brands that need to create unique concepts using AI, or more direct monetization — like paying to access a model or paying for one-off image generation.  For now, though, every little thing on the location is free to make use of. (The corporate uses Cloudflare’s R2 to maintain costs down around downloads.)

“I’d like to start out working with brands and real people and IP to present them the flexibility to sell their likeness to people who need to have the opportunity to make use of it, to do advertisements and things like that,” says Maier. “And I believe that we will arrange licensing in a way and set permissions in a way in order that they will really constrain what it could possibly and can’t be used for. I believe that that’s going to be critical for as we take into consideration how things evolve on this AI era,” he adds.

Over time, the startup goals to expand to other modalities beyond AI image models, but that’s further down the road, Maier says.