About Vectra MCP Server
Vectra MCP Server is a TypeScript server for managing and querying a Vectra knowledge base through its backend API. It provides collection, file, embedding, and hybrid search tools for MCP-compatible clients.
Key features
- 01Create and list collections
- 02Embed text and local files
- 03Manage collection files
- 04Hybrid and graph-enhanced search
- 05Fetch nodes from ArangoDB
- 06Delete files and embeddings
Use cases
- 01Organizing content into Vectra collections
- 02Indexing text and local files
- 03Searching a collection with hybrid search
- 04Inspecting ArangoDB nodes by key
MCP support
It exposes tools to MCP clients for managing Vectra collections and files, embedding content, querying collections, and fetching nodes from ArangoDB.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
We haven’t collected a package or remote endpoint for this server. Follow the instructions in the project README.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMEVectra MCP Server
A Model Context Protocol (MCP) server for interacting with a Vectra knowledge base.
This TypeScript-based MCP server provides tools to manage and query a Vectra instance, enabling integration with MCP-compatible clients. It interacts with a backend Vectra API (presumably running separately).
Features
Tools
This server exposes the following tools for interacting with Vectra:
create_collection: Create a new Vectra collection.- Input:
name(string, required),description(string, optional)
- Input:
list_collections: List existing Vectra collections.- Input: None
embed_texts: Embeds multiple text items in batch into Vectra.- Input:
items(array of objects withtext(required) and optionalmetadata),collectionId(string, optional)
- Input:
embed_files: Reads multiple local files and embeds their content into Vectra.- Input:
sources(array of local file paths, required),collectionId(string, optional),metadata(object, optional - applies to all items)
- Input:
add_file_to_collection: Add an already embedded file (referenced by its ID) to a specific Vectra collection.- Input:
collectionId(string, required),fileId(string, required)
- Input:
list_files_in_collection: List files within a specific Vectra collection.- Input:
collectionId(string, required)
- Input:
query_collection: Query the knowledge base within a specific Vectra collection.- Note: This tool always uses hybrid search (vector + keyword) and enables graph search enhancement by default.
- Input:
collectionId(string, required),queryText(string, required),limit(number, optional),maxDistance(number, optional),graphDepth(number, optional),graphRelationshipTypes(array of strings, optional),includeMetadataFilters(array of objects, optional),excludeMetadataFilters(array of objects, optional)
delete_file: Delete a file and its associated embeddings from Vectra.- Input:
fileId(string, required)
- Input:
get_arangodb_node: Fetch a specific node directly from the underlying ArangoDB database by its key.- Input:
nodeKey(string, required - e.g.,chunk_xyzordoc_abc)
- Input:
(Refer to src/tools.ts for detailed input schemas)
Development
Install dependencies:
npm install
Build the server:
npm run build
Run the server (listens on stdio):
node build/index.js
For development with auto-rebuild:
npm run watch
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.