VectorDb API

Vector DB API

import { CodeBoltClient } from '@codebolt/clientsdk';

const client = new CodeBoltClient();

Quick Reference

Method	Description
addChunks	Adds chunks to a vector collection.
addDocuments	Adds documents to a vector collection.
createCollection	Creates a new vector collection.
createIndex	Creates a vector index for a collection.
deleteChunk	Deletes a specific chunk from a vector collection.
deleteCollection	Deletes a vector collection.
deleteDocuments	Deletes documents from a vector collection.
getChunks	Retrieves chunks from a vector collection.
getCollection	Retrieves a vector collection by its ID.
getDocuments	Retrieves documents from a vector collection.
getSettings	Retrieves settings for a vector collection.
listCollections	Lists all vector collections.
query	Queries a vector collection for similar documents.
updateCollection	Updates a vector collection.
updateSettings	Updates settings for a vector collection.

Methods

---

addChunks

client.vectorDb.addChunks(id: string, data: AddVectorChunksRequest): Promise<unknown>

Adds chunks to a vector collection.

Manually adds pre-chunked text segments to the collection with their embeddings. This is useful when you have custom chunking logic or want to add chunks without full document processing.

Parameter	Type	Required	Description
id	string	Yes	The unique identifier of the target collection
data	AddVectorChunksRequest	Yes	The chunks to add with their embeddings and metadata

Returns: Promise<unknown> — A promise that resolves when chunks have been added

Full reference →

---

addDocuments

client.vectorDb.addDocuments(id: string, data: AddVectorDocumentsRequest): Promise<unknown>

Adds documents to a vector collection.

Ingests new documents into the collection, automatically generating embeddings and storing them for similarity search. Documents can include text content and metadata for filtering.

Parameter	Type	Required	Description
id	string	Yes	The unique identifier of the target collection
data	AddVectorDocumentsRequest	Yes	The documents to add along with their metadata

Returns: Promise<unknown> — A promise that resolves when documents have been added

Full reference →

---

createCollection

client.vectorDb.createCollection(data: CreateVectorCollectionRequest): Promise<VectorCollection>

Creates a new vector collection.

Initializes a new collection with specified configuration including dimensionality, embedding model, and metadata settings. Collections are used to store and manage vectors for similarity search operations.

Parameter	Type	Required	Description
data	CreateVectorCollectionRequest	Yes	The collection configuration data including name, dimensions, and settings

Returns: Promise<VectorCollection> — A promise that resolves to the created VectorCollection object

Full reference →

---

createIndex

client.vectorDb.createIndex(data: CreateVectorIndexRequest): Promise<unknown>

Creates a vector index for a collection.

Initializes an index structure to optimize similarity search performance. Indexes enable fast nearest neighbor queries on high-dimensional vectors and are essential for efficient retrieval.

Parameter	Type	Required	Description
data	CreateVectorIndexRequest	Yes	The index creation configuration

Returns: Promise<unknown> — A promise that resolves when the index has been created

Full reference →

---

deleteChunk

client.vectorDb.deleteChunk(id: string, chunkId: string): Promise<unknown>

Deletes a specific chunk from a vector collection.

Removes a single chunk by its unique identifier. This is useful for granular cleanup when individual chunks need to be removed without affecting other content in the collection.

Parameter	Type	Required	Description
id	string	Yes	The unique identifier of the collection
chunkId	string	Yes	The unique identifier of the chunk to delete

Returns: Promise<unknown> — A promise that resolves when the chunk has been deleted

Full reference →

---

deleteCollection

client.vectorDb.deleteCollection(id: string): Promise<unknown>

Deletes a vector collection.

Permanently removes a collection and all its associated data including documents, chunks, and index information. This operation cannot be undone.

Parameter	Type	Required	Description
id	string	Yes	The unique identifier of the collection to delete

Returns: Promise<unknown> — A promise that resolves when the collection has been deleted

Full reference →

---

deleteDocuments

client.vectorDb.deleteDocuments(id: string, data?: DeleteVectorDocumentsRequest): Promise<unknown>

Deletes documents from a vector collection.

Removes specified documents from the collection based on provided criteria such as document IDs or metadata filters. This operation cleans up the collection and frees storage.

Parameter	Type	Required	Description
id	string	Yes	The unique identifier of the collection
data	DeleteVectorDocumentsRequest	No	Optional deletion criteria for selecting documents to remove

Returns: Promise<unknown> — A promise that resolves when documents have been deleted

Full reference →

---

getChunks

client.vectorDb.getChunks(id: string, params?: Record<string, unknown>): Promise<VectorChunk[]>

Retrieves chunks from a vector collection.

Fetches text chunks that were created when documents were processed. Chunks are smaller segments of documents used for more granular similarity search and retrieval operations.

Parameter	Type	Required	Description
id	string	Yes	The unique identifier of the collection
params	Record<string, unknown>	No	Optional query parameters for filtering or pagination

Returns: Promise<VectorChunk[]> — A promise that resolves to an array of VectorChunk objects

Full reference →

---

getCollection

client.vectorDb.getCollection(id: string): Promise<VectorCollection>

Retrieves a vector collection by its ID.

Fetches detailed information about a specific collection including its configuration, statistics, and metadata. Use this to inspect collection properties or verify collection existence.

Parameter	Type	Required	Description
id	string	Yes	The unique identifier of the collection to retrieve

Returns: Promise<VectorCollection> — A promise that resolves to the VectorCollection object

Full reference →

---

getDocuments

client.vectorDb.getDocuments(id: string, params?: Record<string, unknown>): Promise<VectorDocument[]>

Retrieves documents from a vector collection.

Fetches documents stored in the collection with optional filtering and pagination parameters. Use this to inspect stored content or retrieve specific documents by criteria.

Parameter	Type	Required	Description
id	string	Yes	The unique identifier of the collection
params	Record<string, unknown>	No	Optional query parameters for filtering or pagination

Returns: Promise<VectorDocument[]> — A promise that resolves to an array of VectorDocument objects

Full reference →

---

getSettings

client.vectorDb.getSettings(id: string): Promise<VectorCollectionSettings>

Retrieves settings for a vector collection.

Fetches the configuration and settings applied to a collection including index types, distance metrics, and other operational parameters.

Parameter	Type	Required	Description
id	string	Yes	The unique identifier of the collection

Returns: Promise<VectorCollectionSettings> — A promise that resolves to the VectorCollectionSettings object

Full reference →

---

listCollections

client.vectorDb.listCollections(params?: Record<string, unknown>): Promise<VectorCollection[]>

Lists all vector collections.

Retrieves every collection in the vector database system. Use optional query parameters to filter or paginate results. This is useful for browsing available collections or displaying a collection catalog.

Parameter	Type	Required	Description
params	Record<string, unknown>	No	Optional query parameters for filtering or pagination

Returns: Promise<VectorCollection[]> — A promise that resolves to an array of VectorCollection objects

Full reference →

---

query

client.vectorDb.query(id: string, data: VectorQueryRequest): Promise<VectorSearchResult[]>

Queries a vector collection for similar documents.

Performs semantic similarity search using a query vector or text. Returns documents or chunks ranked by similarity score, enabling retrieval of contextually relevant content for RAG applications.

Parameter	Type	Required	Description
id	string	Yes	The unique identifier of the collection to query
data	VectorQueryRequest	Yes	The query configuration including vector or text and search parameters

Returns: Promise<VectorSearchResult[]> — A promise that resolves to an array of VectorSearchResult objects

Full reference →

---

updateCollection

client.vectorDb.updateCollection(id: string, data: UpdateVectorCollectionRequest): Promise<VectorCollection>

Updates a vector collection.

Modifies the configuration of an existing collection such as name, metadata, or other configurable properties. The collection ID must be valid and the collection must exist.

Parameter	Type	Required	Description
id	string	Yes	The unique identifier of the collection to update
data	UpdateVectorCollectionRequest	Yes	The update data with fields to modify

Returns: Promise<VectorCollection> — A promise that resolves to the updated VectorCollection object

Full reference →

---

updateSettings

client.vectorDb.updateSettings(id: string, data: UpdateVectorSettingsRequest): Promise<VectorCollectionSettings>

Updates settings for a vector collection.

Modifies the configuration parameters of a collection such as distance metric, index type, or other settings. Some settings may require index rebuild or have performance implications.

Parameter	Type	Required	Description
id	string	Yes	The unique identifier of the collection to update
data	UpdateVectorSettingsRequest	Yes	The settings to update with new values

Returns: Promise<VectorCollectionSettings> — A promise that resolves to the updated VectorCollectionSettings object

Full reference →