AI image generator Strong Diffusion there are already many fans, and now those experimenting with new AI technology to develop their own models have a place to share their work with other enthusiasts. A start called Civitai — a play on words Civitasmeans community — has created a platform where members can post their own Stable Diffusion-based AI image models for others to discover, as well as the output of their work — AI photos — for consumers to browse and enjoy.
EXPLAIN Civitai CEO Justin Maier, the idea for the startup happened because he recognized that there is a need for a place where people can share their models and others can see them. People, he said, will post the images they have made but some don’t know how to duplicate their work to make their own images.
After a project at Microsoft, where he worked as a contractor on web development projects, Maier found himself interested in Midjourney.
“It captures the same itch that web development has where I see something – or I have something in mind – and I can type text to get something out. It becomes a collaborative experience where I can explore my creativity and be drawn in directions I might not have planned,” he said.
But Maier soon found himself so limited to Midjourney’s credit-based plan that he ended up having to wait 5 minutes for each image to be created. That, he said, was “a real bummer because I really fell in love with this ability to explore and create anything I could think of,” Maier told TechCrunch.
At the same time, an open-source version came out called Stable Diffusion which basically allows you to do the same thing in terms of AI image generation. As the community around Stable Diffusion grows, people learn how to create different things in the model by prompting or by adding new concepts to the model, such as using images in their itself to create AI-generated selfies, for example. Maier started to see people posting their own models that would allow you to do cool things – like create a synth wave style or punk style – and combine the models to create other new concepts as well. But it’s posted on the web in places like Reddit or Discord, not in a centralized community.
That led him to create Civitai (or Model Share, as it was originally called), an online community that organizes this content in a way where people can come back and find it later.
Initially, he seeded Civitai by contacting model creators and asking them if they could post their work on the site. Initially, there were about 40 or 50 different models available. But over time, the site began to grow.
“People are happy to see this effort that they’ve made. And so around January, we became a place for sharing these things,” Maier said. “There are other sites that post about human, like HuggingFace…but it’s not image-centric,” he explained. “So we’ve become the de facto standard for sharing your model and various AI resources and images you’ve created,” Maier added.
By January 2023, the site hit 100,000 registered users and it occurred to Maier that it could be more than a community project – it could be a company. So Civitai was once “officially” built, and three months later, it hit one million registered users. Today, that number is close to 3 million registered users and it sees about 12 to 13 million unique visitors every month.
“It’s grown so fast — definitely bigger than we thought it would be back in January,” Maier said.
Of the 3 million users, only about 10,000 unique creators each month actually upload new models. However, that number increased by approximately 25% last month, after Citivai added the ability for people to train new models on the site, making it easier to get started. Many users use the models and the content created.
As a result of its development, Civitai, also founded by Maxfield Hulker and Briant Diehl, raised a $ 5.1 million round in June led by Andreessen Horowitz (a16z) at a $ 20 million value . Some of the fund’s participants are a bit unusual – the legal team that helped them close the round is also on board.
“Civitai is a prime example of a company that has built an incredible, engaged community and all without spending a fortune on marketing,” Andreessen Horowitz Partner Brian Kim said in a statement. . “Our investment in the company will only supercharge something that is already working so well, by contributing to a world where any individual can take advantage of what the biggest technological change in our generation – AI.”
To use Civitai, users can upload a series of images that represent the style represented by the images, then select a base model — such as a standard model, an anime model, or more realistic model, etc. After about an hour, the new model should be ready and you will be able to create your own pictures using the style you got. Images generated on the site include metadata detailing things like prompts and resources used, and Civitai encourages off-site modelers to do the same.
There are issues with artists knowing their work is being used to train AI models, which is a concern. To address this, Civitai created a process that allows artists to flag the resources they believe their work uses, which starts a negotiation on next steps. In some cases, that’s the complete removal of the resource, or in other cases, the artist just wants their name removed.
“Really what we want to happen – ultimately – is that these artists can … use these styles for themselves. And if they want to give people the ability to create their style, then they can give people the ability to pay to do that,” Maier said. “But it’s early days and I haven’t had a chance to work with a lot of artists who are interested in doing that,” he admits.
However, Maier believes there will be a market for that kind of work in the future, which will include the use of AI imagery within things like movies, music videos, apps, and other creative endeavors. . In the near term, however, the plan is to create a consumer-facing mobile app that will act as an AI image repository, as a kind of companion to the experience taking place on the main site. .
The company also now plans to focus on allowing users to monetize their work, whether that’s connecting them with brands that want to create unique concepts using AI, or more direct monetization – like to pay to access a model or pay for a generation image. . For now, however, everything on the site is free to use. (The company uses Cloudflare’s R2 to reduce the cost of downloads.)
“I want to start working with brands and real people and IP to give them the ability to sell their likeness to people who want to use it, make ads and things like that,” said Maier. “And I think that we can set up licensing in a way and set permissions in a way so that they can really restrict what can and can’t be used. I think that’s going to be critical as we figure out how the things that have evolved in this era of AI,” he added.
Over time, the startup aims to expand into other modalities beyond AI image models, but that’s further down the road, Maier said.