vectra-mcp-server
If you are the rightful owner of vectra-mcp-server and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.
A Model Context Protocol (MCP) server for interacting with a Vectra knowledge base.
The Vectra MCP Server is a TypeScript-based server designed to facilitate interaction with a Vectra knowledge base. It provides a suite of tools for managing and querying Vectra instances, allowing seamless integration with MCP-compatible clients. The server communicates with a backend Vectra API, which is assumed to be running separately. This server is particularly useful for creating and managing collections, embedding texts and files, and querying the knowledge base using advanced search techniques. It supports hybrid search capabilities, combining vector and keyword searches, and enhances queries with graph search functionalities.
Features
- Create and manage Vectra collections.
- Embed texts and files into the Vectra knowledge base.
- Query collections using hybrid search and graph search enhancements.
- Delete files and their embeddings from Vectra.
- Fetch nodes directly from the underlying ArangoDB database.
Tools
create_collection
Create a new Vectra collection.
list_collections
List existing Vectra collections.
embed_texts
Embeds multiple text items in batch into Vectra.
embed_files
Reads multiple local files and embeds their content into Vectra.
add_file_to_collection
Add an already embedded file to a specific Vectra collection.
list_files_in_collection
List files within a specific Vectra collection.
query_collection
Query the knowledge base within a specific Vectra collection.
delete_file
Delete a file and its associated embeddings from Vectra.
get_arangodb_node
Fetch a specific node directly from the underlying ArangoDB database by its key.