Loci API
Caption 3D Scene
Returns a descriptive caption generated for a 3D scene.
*Help: Generally, a scene will consist of an environment or landscape, built up of constituent objects (an object being a singular, isolated entity). Examples of scenes include a room layout, street scene or woodland.*
#### Accepted Scene File Formats:
- .3DS
- .FBX
- .GLB
- .GLTF
- .MAX
- .OBJ
- .OFF
- .PLY
- .STL
- .USDZ
or a .ZIP file with exactly one scene file of one of the above formats.
#### Adding Textures:
Textures/Materials are important for caption accuracy, and so care must be taken to ensure that they are either:
1. Embedded within the single scene file
2. Referenced by relative path in the scene file. If your scene uses separate (ie. non-embedded) texture files you must first ZIP up the folder containing the scene file and textures.
###
*Hint: As a rule of thumb, if your scene's textures correctly load in Blender, then they will correctly load in the API.*
###
You can check whether textures have been applied properly by looking at the rendered image of your scene in the API Dashboard
Caption 3D Asset
Returns a descriptive caption generated for a 3D asset.
#### Accepted Asset File Formats:
- .3DS
- .FBX
- .GLB
- .GLTF
- .MAX
- .OBJ
- .OFF
- .PLY
- .STL
- .USDZ
or a .ZIP file with exactly one asset file of one of the above formats.
#### Adding Textures:
Textures/Materials are important for caption accuracy, and so care must be taken to ensure that they are either:
1. Embedded within the single asset file
2. Referenced by relative path in the asset file. If your asset uses separate (ie. non-embedded) texture files you must first ZIP up the folder containing the asset file and textures.
###
*Hint: As a rule of thumb, if your asset's textures correctly load in Blender, then they will correctly load in the API.*
###
You can check whether textures have been applied properly by looking at the rendered image of your asset in the API Dashboard
Classify 3D Asset
Returns a classification for a 3D asset according to a taxonomy.
Available Loci Taxonomies:
- [`loci-object`](https://loci-public.s3.eu-west-2.amazonaws.com/taxonomies/loci-object.json) (default)
3D Asset Color Classification
Determines the color composition of a 3D asset.
Embed a 3D Asset using Loci's 3D Encoder
Returns an embedding for a 3D asset which can be used in combination with the /image/embed and /text/embed endpoints
for a variety of downstream tasks such as 3D similarity search, zero-shot classification, semantic text search, reverse image search, and more.
For a managed service for 3D similarity search, semantic text search and reverse image search, 3D recommendations, see our 3D search API endpoints (/3d/index, /3d/find-similar).
For a managed service for zero-shot 3D classification, see our 3D classification API endpoint (/3d/classify), or contact our CEO at eugene@loci.ai for custom taxonomies.
Available Loci Encoders:
- [`loci`] (default)
3D Asset Similarity Search
Search for similar or duplicate 3D assets.
Get paginated 3D Assets in the search index
Retrieve the 3D assets from the search index.
Index a 3D Asset
Index a 3D asset to the search index. Note that indexing an asset will overwrite any existing asset with the same asset ID.
Delete an entire search index
Delete the entire search index
Get Information about your search index
Retrieve the 3D assets from the search index.
Delete an asset in your search index by ID
Delete an 3D asset from the search index by providing the asset_id.
Metadata of a 3D Asset
Returns Technical Metadata for a 3D asset.
- Number of vertices
- Number of edges
- Number of polygons
- Number of submeshes
- Number of animations
- Animation names
- Number of materials
- Material names
- Number of texture images
- Texture names
- Mesh names
- Is rigged
- Dimensions
Optimize 3D Asset
Optimize a 3D Asset for the web by converting it to a .glb with compressed textures and Draco-compressed geometry.
Render of a 3D Asset
Returns a render of a 3D asset with automated camera positioning and lighting.
Tag 3D Asset
Returns free form tags generated for a 3D asset.
#### Accepted Asset File Formats:
- .3DS
- .FBX
- .GLB
- .GLTF
- .MAX
- .OBJ
- .OFF
- .PLY
- .STL
- .USDZ
or a .ZIP file with exactly one asset file of one of the above formats.
#### Adding Textures:
Textures/Materials are important for tagging accuracy, and so care must be taken to ensure that they are either:
1. Embedded within the single asset file
2. Referenced by relative path in the asset file. If your asset uses separate (ie. non-embedded) texture files you must first ZIP up the folder containing the asset file and textures. You can read more about how to pack textures [here](https://docs.loci.ai/docs/limits-limitations).
###
*Hint: As a rule of thumb, if your asset's textures correctly load in Blender, then they will correctly load in the API.*
###
You can check whether textures have been applied properly by looking at the rendered image of your asset in the API Dashboard
Embed an Image using Loci's Image Encoder
Returns an embedding for an image which can be used in combination with the /3d/embed and /text/embed endpoints
for a variety of downstream tasks such as 3D similarity search, zero-shot classification, semantic text search, reverse image search, and more.
For a managed service for 3D similarity search, semantic text search and reverse image search, 3D recommendations, see our 3D search API endpoints (/3d/index, /3d/find-similar).
For a managed service for zero-shot 3D classification, see our 3D classification API endpoint (/3d/classify), or contact our CEO at eugene@loci.ai for custom taxonomies.
Available Loci Encoders:
- [`loci`] (default)
Tag Image
Returns free form tags generated for an image.
#### Accepted Image File Formats:
- .PNG
- .JPEG
- .JPG
- .WEBP
Embed text using Loci's 3D/image/text Encoder
Returns an embedding for text which can be used in combination with the /image/embed and /3d/embed endpoints
for a variety of downstream tasks such as zero-shot classification, semantic text search. The embeddings from this
/text/embed endpoint are trained to be semantically aligned with the embeddings from the /image/embed and /3d/embed endpoints.
For a managed service for 3D similarity search, semantic text search and reverse image search, 3D recommendations, see our 3D search API endpoints (/3d/index, /3d/find-similar).
For a managed service for zero-shot 3D classification, see our 3D classification API endpoint (/3d/classify), or contact our CEO at eugene@loci.ai for custom taxonomies.
Available Loci Encoders:
- [`loci`] (default)
Get all available workspaces
Get all available workspaces
Create workspace
Create workspace
Get information about a workspace from it's ID
Get information about a workspace from it's ID
Update a workspace by Invite user(s)
Update a workspace by Invite user(s)