Foveate Docs

AI Content and Workflows

Foveate is designed to be the sharing and storytelling layer for interactive 3D experiences and accepts any type of content whether its human or AI designed. Mix and match all sorts of content media a
AI is quickly revolutionizing the content generation workflow in every media category. Using AI tools can exponentially speed up your workflow if you are an experienced creator or help novices produce media they otherwise would not be able to create.
While we appreciate the advancements in AI technology it is important to maintain ethics with such power. At Foveate, we will never train an AI on your work without your consent.

3D Content Creation

Currently, 3D has the least developed of AI tools as 3D content is notoriously difficult to create for human authors and the datasets available to train AI models on 3D are not as large or as curated. We will cover some of the most promising tools.
Explore the possibilities of 3D creation with AI, 3Dpresso can extract the 3D model of your object from a 1-minute video. 3DPresso is currently the best and most impressive of the AI-generated 3D model platforms. It works amazingly well for small and medium-sized objects and since the input data is image-based your results are highly predictable. 3Dpresso has an excellent Video to 3D pipeline and also boasts some other workflows like text to 3D as well, albeit those models are currently quite limited to the datasets they are trained on. Learn more here.
Common Sense Machines (CSM)
Create 3D assets from video, image, or text. Generate assets with high-resolution geometry, UV-unwrapped textures, and neural radiance fields using the latest breakthroughs in neural inverse graphics. Now creating environments and games is faster and more accurate than ever before. CSM can take a single image and hallucinate a 3D model from it that is game-ready. This is incredible for creating prototypes very quickly. Learn more here.
Capture in lifelike 3D. Unmatched photorealism, reflections, and details. Luma is a Neural Radiance Field-based app, meaning that it is not specifically designed to capture and produce 3D models, it is more geared towards producing video output that accurately portrays the lighting and reflection in a captured scene, but can still output a 3D model of an object or space. The meshes it produces are not optimized for 3D workflows and will require cleanup in another app to look correct in a real-time experience. Luma does offer an Unreal plugin which allows the native rendering of NeRFs. They are not suitable for gameplay but can be used to render Cinematic quality videos. Learn more here.

2D Content Creation

The options for 2D image creation with AI is now endless, but here are some of our favorite apps and recommendations for how to best use them for different applications.
Midjourney is an AI text-to-image generation program that is run off of Discord. Learn more here.
WithPoly is an AI text-to-image generation program designed specifically for texture generation for 3D workflows. It can produce all the texture maps necessary for a PBR (Photorealistic Based Rendering) workflow and allow you to make your textures seamless so that they can be tiled over your 3D model. Learn more here.

Text Content Creation

ChatGPT is an artificial intelligence chatbot developed by OpenAI that can be very useful for finding information or for writing text. Use it to write concise captions for your Moments. It is also very good at creating prompts for other AI tools, for example, ask it to write a prompt tailored for Midjourney. Learn more here.
Last modified 1mo ago