I’m so excited to launch this!
It started with me and @larsvagnes experimenting with deep-fakes, and we were wondering if we could make a useful product with it.
We figured we could use the technology to give people the ability to create engaging videos really fast.
Our prototype started as a simple text-to-video (video of the generated human saying it), but we built it into a more full-fledged studio for creating simple marketing videos.
We’re not completely sure what our main use-case will be, will it be people who want to create a video fast? or people who would like their video localized in multiple languages? We hope we’ll get some validation and pointers on the future after this launch.
Another exciting possibility this technology opens is that you can batch send personalized videos to people on a list. We are doing a pilot project with a company with a newsletter with 20K people, where we’ll AB-test with a personalized video for half of them. Our guess (or hope?) is that the personalized video will yield high conversion rates and engagement.
You can preview and test with as many videos as you would like. We charge a price for rendering and exporting the video. (it takes around 15 minutes).
PS: The model’s face will not move in the studio, but will move human-like when exporting!
Let us know what you think, we are ready to answer questions and fix bugs!
– Lars Karbo