Vectra MCP Server
by theVuArena
The Vectra MCP Server is a TypeScript-based server designed to interact with a Vectra knowledge base using the Model Context Protocol (MCP). It provides tools to manage and query a Vectra instance, enabling integration with MCP-compatible clients.
Last updated: N/A
What is Vectra MCP Server?
The Vectra MCP Server is a Model Context Protocol server built in TypeScript that facilitates interaction with a Vectra knowledge base. It acts as an intermediary between MCP-compatible clients and the Vectra API, providing tools for managing collections, embedding data, and querying the knowledge base.
How to use Vectra MCP Server?
To use the server, first install dependencies using npm install, then build the server with npm run build. Run the server using node build/index.js. The server listens on stdio. For development, use npm run watch for auto-rebuild. The server exposes various tools, each with specific input requirements as detailed in src/tools.ts.
Key features of Vectra MCP Server
Create and manage Vectra collections
Embed text and files into Vectra
Query Vectra collections using hybrid search
Delete files and associated embeddings
Fetch specific nodes from the underlying ArangoDB database
Use cases of Vectra MCP Server
Integrating Vectra knowledge base with MCP-compatible applications
Building custom knowledge management solutions
Developing AI-powered applications that require access to a knowledge base
Programmatically managing and querying a Vectra instance
FAQ from Vectra MCP Server
What is Vectra?
What is Vectra?
Vectra is a knowledge base system (details not provided in the document).
What is MCP?
What is MCP?
MCP stands for Model Context Protocol.
Where can I find the input schemas for the tools?
Where can I find the input schemas for the tools?
The input schemas for the tools are detailed in src/tools.ts.
How do I run the server in development mode?
How do I run the server in development mode?
Use the command npm run watch for development with auto-rebuild.
What type of search is used for querying?
What type of search is used for querying?
The query_collection tool always uses hybrid search (vector + keyword) and enables graph search enhancement by default.