Member Junction
    Preparing search index...

    Provides a simple interface for InteractiveComponents to perform a wide variety of common AI operations such as prompt execution with LLMs, calculating embeddings on strings, and using vector search for small to mediun sized datasets in memory.

    interface SimpleAITools {
        EmbedText: (
            params: SimpleEmbedTextParams,
        ) => Promise<SimpleEmbedTextResult>;
        ExecutePrompt: (
            params: SimpleExecutePromptParams,
        ) => Promise<SimpleExecutePromptResult>;
        VectorService: SimpleVectorService;
    }
    Index

    Properties

    EmbedText: (params: SimpleEmbedTextParams) => Promise<SimpleEmbedTextResult>

    Used to calculate vector embeddings for one or more strings. Uses very fast small/medium sized local models so the embeddings can be rapidly calculated for hundreds or even thousands of pieces of text. This allows interactive components to dynamically compute similarity/distance between any kinds of data and generate very interesting interactive experiences for users

    Type Declaration

    ExecutePrompt: (
        params: SimpleExecutePromptParams,
    ) => Promise<SimpleExecutePromptResult>

    Uses an LLM to respond to a provided prompt. Often used by interactive components to provide rich analysis of data within a component that a user is interested in gaining qualitative insights on

    Type Declaration

    VectorService: SimpleVectorService

    Instance of the SimpleVectorService that can be used by Interactive Components

    SimpleVectorService for more details on this. This object can perform a wide array of vector data operations such as KNN, Similarity Scoring, and more.