Loci
Loci uses Machine Learning to tag, search, and recommend 3D assets— so your team can spend less time sorting, and more time creating.
Loci is a platform for fast and seamless creation of virtual 3D worlds.
We enable 3D artists, businesses and individuals to create immersive 3D virtual world replicas of any real-life image, using state-of-the-art computer vision models in image-based 3D asset search.
Artificial intelligence
for the third dimension.
Tag, search and discover assets using the
most accurate machine learning platform for 3D.
Loci uses Machine Learning to tag, search, and recommend 3D assets—
so your team can spend less time sorting, and more time creating.
Automated 3D Asset Tagging
We automatically tag your 3D assets with completely flexible taxonomies.
Neural Search Engine for 3D Assets
The most accurate search engine for 3D assets. We understand the context of your search, without reliance on manual tags and return the most relevant results.
Visual Search for 3D Assets
Search for 3D assets by uploading an image of the object you're looking for.
How does Loci AutoTag work?
Loci AutoTag uses a combination of state-of-the-art computer vision and natural language processing techniques to automatically tag your 3D assets. We encode all the visual information in the asset, as well as any textual information from the asset name, or pre-existing tags using neural networks. Next, given a completely flexible taxonomy of tag categories (e.g. “Color”, “Material”, “Style”, “Biome”, etc.), our model generates the most relevant tags for each category.
Can we use Loci AutoTag with our own taxonomy?
Yes! We have suggested taxonomies for you to use, but you can also create your own.
How do we integrate Loci AutoTag with our existing pipeline?
You can integrate Loci AutoTag with your existing pipeline by either using our export functionalities on our front-end dashboard, or by using our API. We have a number of different endpoints that you can use to index your 3D assets, setup your taxonomies, and retrieve the tags.
How does Loci Search work?
We encode each asset's visual information by passing it through our machine learning encoder which has been trained on hundreds of millions of datapoints. So when a user performs a text or image-based search query with Loci Search, we search based on how the 3D asset actually looks, rather than having to rely on the manual tags or asset name.
What kind of search queries does Loci Search support?
Loci can handle descriptive text, synonyms, and typos - all in 50+ languages. In addition, our visual search can handle images of individual objects, cropped scenes, and stylized sketches.
How do we integrate Loci Search with our existing pipeline?
You can integrate Loci Search with your existing pipeline by using our API. Loci Search is also built into our UI dashboard for free, so you can try and test our neural 3D asset search on your library.
