Genmo Chat is the first step towards our approach to Creative General Intelligence where a human and a generative model work together. Today, Genmo Chat can help you create images, videos and 3D models. Collaboration yields more creative and useful results than any AI alone.
Generative models have demonstrated incredible capabilities in synthesizing content across modalities, including text, images, videos, and beyond.
With Genmo, we're taking it a step further by providing a creative copilot that works hand-in-hand with users to bring their creative visions to life.
We are gradually rolling out alpha access to Genmo Chat to creatives from our waitlist. While there are limitations, we are scaling up capacity and continuously working to improve Genmo's capabilities, safety, and understanding of user intent.
Genmo can animate existing images. The user uploads a starry night and asks Genmo to animate the sky into a timelapse. The user controls the animation by asking Genmo to only animate the sky and not the mountain.
Genmo can generate and edit movies from scratch. The user asks Genmo to create a movie with a title. The model will help create ideas which the human can critique iteratively. Genmo takes it from there to generate an edited video.
Genmo opted to use our V2 video generation model because it can generate coherent global motion. It also automatically selects transitions and text overlays to match the plotline.
Like the previous example, Genmo can generate and edit movies from scratch. The user asks to generate a movie called “Godfather: The Lunar Family”.
Genmo helps the user refine their ideas into a proposed script. Genmo generates a variety of scenes and transitions. In this example, the user works with Genmo to create a poster photo.
Replace content and change image styles with natural language. Genmo allows users to direct the creative process at a high level, while the model suggests specific details and calls the necessary tools to get the job done.
Expect even higher image quality today. The demo uses our old Genmo V2 model, and we&'ve since upgraded to a new V3 image generator.
Genmo can generate app icons as well. Here, Genmo makes icons for a “creative copilot”.
In response to user feedback, Genmo regenerates variations of their favorite icon. Finally, Genmo combines all the images into a slide deck to share with the team.
To bridge the gap between humans and generative tools, we're working on improving our models' understanding of user intent and context. This will allow for more seamless collaboration between users and their creative copilot, ultimately leading to better and more useful results.
As a creative assistant, Genmo supports a wide range of tools, such as text-to-image, image editing, image enhancement, video generation, and more. By using natural language, users can instruct Genmo to perform various tasks, including generating new images from descriptions, editing existing images, or even creating looping videos.
We believe that collaboration is the missing piece from current generative AI models. We are building Genmo to transform the way we as people create content across modalities. Here's what we envision for the future:
Our beta users have found the prototype of Genmo Chat to be useful for tasks across marketing, content creation, design and also just for fun. However, there are some limitations in the current version of Genmo. For example, Genmo may not always understand user intent, especially when ambiguous. The agent is constantly improving with user feedback.
We have ambitious goals for Genmo, and we know that achieving them requires a collective effort. We invite youto join us in exploring the immense potential of generative models. If you're excited about the future we're building and want to be a part of it, please reach out.
Looking for Genmo V1? The classic experience is still at alpha.genmo.ai