Semantic functions (preview)
Semantic functions provide solution partners with a fast and easy way to integrate AI capabilities into their solutions without requiring DAM, PIM, or other Bizzkit applications.
Info
Please note that semantic functions are in preview. Read more here.
Several endpoints are available on the BAIA preview API:
- ImageDescription: Endpoints related to extracting descriptions from images, such as identifying objects, scenes, and activities.
- ImageSearchableData: Endpoints related to extracting searchable data (themes, colors, categories, etc.) from images to enhance search capabilities.
- IntentEmbedding: Endpoints related to extracting intents from text, useful for understanding user queries and commands.
- ProductCategoryPlacement: Endpoints related to categorizing products into appropriate categories based on their attributes.
- ProductChoice: Endpoints related to recommending products based on user preferences and behavior.
- ProductEmbedding: Endpoints related to generating product embeddings for similarity searches and recommendations.
- ProductTextGeneration: Endpoints related to generating product descriptions, titles, and other textual content.
- SearchSpecialization: Endpoints related to search specialization where context is important, improving search relevance.
- SemanticTranslation: Endpoints related to translating text while preserving its meaning and context.
- TypeAhead: Endpoints related to type-ahead functionality, providing real-time suggestions as users type.
Tip
The endpoints primarily fall into two categories: asynchronous functions (which include start/bulk and status endpoints) and synchronous functions (typically referred to as call-and-wait).
Examples of semantic functions
Here are some simple examples of how to use some of the semantic functions to gain a better understanding of the API.
Image description
The ImageDescription endpoint is used to extract descriptions from images. The following request example shows how to extract descriptions from an image:
Based on an image like this:
the result could be:
Image searchable data
The ImageSearchableData endpoint is used to extract searchable data from images. The following request example shows how to extract searchable data from an image:
Based on an image like this:
the result could be:
Translations
The SemanticTranslation endpoint is used to translate text while preserving its meaning and context. The following request example shows how to translate text:
The result could be: