home / mcp / vectorize mcp server

Vectorize MCP Server

Official Vectorize MCP Server

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "vectorize-io-vectorize-mcp-server": {
      "command": "npx",
      "args": [
        "-y",
        "@vectorize-io/vectorize-mcp-server@latest"
      ],
      "env": {
        "VECTORIZE_TOKEN": "${input:token}",
        "VECTORIZE_ORG_ID": "${input:org_id}",
        "VECTORIZE_PIPELINE_ID": "${input:pipeline_id}"
      }
    }
  }
}

You deploy the Vectorize MCP Server to connect Vectorize’s advanced vector search and text extraction capabilities with your MCP workflows. This server runs locally or remotely and exposes the Vectorize features to your MCP client, enabling powerful document retrieval, extraction, and deep research tasks.

How to use

You interact with the Vectorize MCP Server through an MCP client by configuring a local or remote MCP connection. The server provides three core capabilities: vector-based document retrieval, text extraction with chunking into Markdown, and private deep research generation from your pipeline. Use these tools to search your knowledge base, extract structured text from documents, and run focused research workflows. Start by ensuring your MCP client is pointed to the Vectorize MCP Server configuration described in your workspace, then invoke the available functions as you would with any MCP endpoint: retrieve documents by a query, extract content from uploads, and generate deep research outputs.

How to install

Prerequisites: you need Node.js and npm installed on your system. Verify installation with node -v and npm -v.

Set up your Vectorize credentials and pipeline identifier before starting the server. You will provide these values to the server via environment variables.

export VECTORIZE_ORG_ID=YOUR_ORG_ID
export VECTORIZE_TOKEN=YOUR_TOKEN
export VECTORIZE_PIPELINE_ID=YOUR_PIPELINE_ID

npx -y @vectorize-io/vectorize-mcp-server@latest

Additional sections

Configuration and environment variables for the Vectorize MCP Server are shown below. The server runs as a local process and is started with the npx command, using your Vectorize credentials to initialize the connection.

Environment variables used by the server include the following. Replace the placeholder values with your actual Vectorize details.

{
  " VECTORIZE_ORG_ID": "YOUR_ORG_ID",
  " VECTORIZE_TOKEN": "YOUR_TOKEN",
  " VECTORIZE_PIPELINE_ID": "YOUR_PIPELINE_ID"
}

Development notes

From development and release perspectives, this server is driven via npm scripts. You can install dependencies, build if needed, and run development commands as described in typical Node.js workflows.

Available tools

retrieve

Perform a vector search to find relevant documents by a query and return the top k results.

extract

Extract text from a document (such as a PDF) encoded as base64 and chunk it into Markdown-ready sections.

deep-research

Create a Private Deep Research from your pipeline with an optional web search to gather contextual information.