Create a SemanticSearchIndex:

Extract the most relevant quote from Alice in Wonderland:

Create a SemanticSearchIndex with a single source:

Find index items closest to a given query:

Get a list of available result properties:

Extract a specific item property:

Get multiple properties:

Retrieve all available properties:

Define an index with labeled sources:

The label is automatically returned when searching:

Return the item instead:

Return the item and the label:

Filter to only search items with a specific label:

Define an index with tagged sources:

The tags are automatically returned when searching:

Return a specific tag value:

Return multiple tags:

Return tags and other properties:

Filter for a specific tag value:

Search Shakespeare's sonnets using partial quotes. Use a manual splitting of the text to group each sonnet together with its number:

Then label each sonnet with its number:

Create a index using the labeled sonnets:

Retrieve sonnets and their number using a partial quote:

Look for a vaguely remembered sonnet:

Retrieve the index of the Wolfram Function Repository:

Search for custom functions that can generate a graphic of the solar system:

Search for a specific example in the given function:

Run the example function from the given item:

Create an index from a set of books:

Find some examples on a topic of interest:

Use LLMSynthesize to give a summary contrasting the differences:

Create an index containing German and English text:

The embeddings of the same meanings are similar across languages. Search in English for information in the German source:

Translate the result into English:

When performing a search, the distance function specified in CreateSemanticSearchIndex is used:

The distance function setting can be retrieved from the VectorDatabaseObject backing the index:

Create a index with labels:

Items will always share the tag of their source:

To get list of source tags, use unique tags and DeleteDuplicates:

SemanticSearch automatically reranks the results using SemanticRanking:

This is the equivalent of retrieving 10 results without reranking (the default "MinItems" for reranking):

Then taking the first five results of the results after using SemanticRanking: