Stable Diffusion Synthetic Media Hub Civitai Raises $5.1M
Synthetic media content marketplace Civitai has raised $5.1 million in a seed funding round led by Andreessen Horowitz. Civitai offers a digital hub for images and models produced through the open-source generative AI model Stable Diffusion, where creators can share their output and eventually even license their variant models.
Civitai allows members to post their own Stable Diffusion AI image models and allow others to play around with them and see what comes of them. Users can also upload example images representing a desired style to train custom AI models tailored to their aesthetic. After about an hour of processing, the new models are ready to generate original images.
Civitai encourages detail-rich metadata to properly credit the prompts and resources used. However, controversies have arisen around art being included without artists’ consent. Civitai established a copyright process for removing problematic resources. Civitai has also made a point of highlighting its restrictions against suggestive content and on-site image generation of real individuals without permission. It also noted only 10% of user “bounties” involve actual people’s likenesses.
“With the proliferation of generative AI advancements driven by companies like Stability AI, Midjourney, and OpenAI, people are seeking out new ways to collaborate, share, and learn about GenAI – and they’ve landed on Civitai to do just that, “Civitai CEO Justin Maier explained in a blog post. “With a strong foundation in place, we’ll then work to build upon everything that made our community and platform the special place it is today. Generative AI presents so much potential for creators and builders, and we’re excited to be a part of helping unlock it, every step of the way!”
Civitai boasts 10 million monthly visitors and three million registered members, with around 7,000 new members joining each day. While currently focused on Stable Diffusion, Civitai ultimately aims to support multi-modal generative AI spanning text, images, video, and audio. By year’s end, new features like revamped profiles, messaging, and an on-site image generator will lay the “v1” foundation, according to Maier.