Get the number of vectors in the index

Clone the cached client (shares the same cache and statistics)

Check if the index is empty

Check if a vector exists

Get a service by type

Returns the update ID as a string slice

Check if a timestamp falls within this range

Get the number of registered services

Get all entities

Finish the timer and return the elapsed duration

Set value in cache Stores in both L1 (in-memory) and L2 (Redis) for redundancy. # Arguments `key` - Cache key `value` - Value to cache

Get all chunks

Add an entity to the knowledge graph

Add a document to the knowledge graph # Arguments `id` - Unique document identifier `text` - Document text content `embedding` - Pre-computed embeddi

Check if the index is empty

Check if cluster is empty

Add a node to the incremental graph

Get the next event

Remove a service

Add an error and mark validation as failed

Store a value in an object store # Arguments `store_name` - Name of the object store `key` - Key to store the value under `value` - Value to store (m

Get all relationships in the graph

Build the vector index Must be called after adding documents and before querying.

Check if a file exists

Add a relationship between entities

Add a mapping between an entity and a chunk This is idempotent - adding the same mapping multiple times has no effect

Build the GraphRAG instance This method is only available when both output directory and LLM/embedding backend are configured. # Example ```no_run u

Check if the registry is empty

Query the knowledge graph # Arguments `query_embedding` - Pre-computed query embedding `top_k` - Number of results to return # Returns JSON string w

* Sleep utility for yielding control * @param {number} ms * @returns {Promise<void>}

Add temporal edge

Clear all data

Get a mutable service by type

Add a grounded entity to this anchor

Add a relationship to this cluster

Add a document from text content

Split text into chunks with overlap (legacy method)

Generate embedding for a text string

Execute a logic form query against the knowledge graph

Set content from plain string lines (runs through markdown parser)

Embed a single text

Add a vector to the index

Simple chat with a single user message # Arguments `user_message` - User's message # Returns Assistant's response

Add a warning (doesn't affect validity)

Get all relationships as an iterator

Initialize the async GraphRAG system

Load model from Cache API Downloads model files from HuggingFace if not cached. Call this method before calling embed() or embed_batch().

Add concepts from a document

Chunk text according to strategy

Async version of complete

Generate a test document with realistic content

Get an entity by ID

Get a reference to the knowledge graph

Execute comprehensive search using both vector similarity and PageRank

Get entity details by ID

Get neighbors of an entity with their connecting relationships

Load the index from IndexedDB

Register a service implementation

Run the layout algorithm for a fixed number of iterations

Register a service and continue building

Add an optimization suggestion

Build a knowledge graph from a collection of documents

Send a chat message and get response # Arguments `messages` - Array of chat messages `temperature` - Sampling temperature (0.0-2.0, default 0.7) `max

Cosine similarity

Create an entity with mentions

Get all documents

Estimate cost of an operation

Extract sentences from text

Get style for normal text

Add an embedding to the index

Add concepts from a chunk

Save the index to IndexedDB for persistence

Set focused pane and update all component focus states. pane: 0=input, 1=results, 2=raw_results, 3=info_panel

Build the concept graph

Calculate coherence score for a chunk (based on word overlap)

Clean word by removing punctuation

Extract entities from a text chunk using dynamic entity types

Get a connection to the Redis server

Initialize GraphRAG with configuration

Initialize vector index from the knowledge graph

Set the output directory (required) This transitions the builder to the `HasOutput` state.

Add a chunk to the knowledge graph

Add a node to the layout

Add a custom synonym mapping # Arguments `term` - The original term `synonyms` - List of synonyms