Overview

The Supavec MCP Server bridges AI models with vector databases, allowing for efficient storage, retrieval, and management of high-dimensional vector embeddings. This is crucial for AI applications such as Retrieval Augmented Generation (RAG), semantic search, and recommendation systems, enabling AI to understand and interact with data based on its meaning and context. For more details, explore its Key Features, learn about Available Tools, and find Installation instructions. You can also see Common Use Cases.

Official Server:

Developed and maintained by Model Context Protocol

Key Features

💾

Vector Storage & Indexing

Efficiently store and index high-dimensional vector embeddings for fast retrieval.

🔍

Semantic Search

Perform similarity searches using natural language queries or vector inputs to find relevant data.

🗂️

Metadata Filtering

Refine search results by applying filters based on associated metadata with your vectors.

🚀

Scalability

Designed to handle and scale with large-scale vector datasets and high query loads.

Available Tools

Quick Reference

Tool	Purpose	Category
`supavec_upsert`	Insert or update vectors with metadata	Write
`supavec_query`	Perform a similarity search	Read
`supavec_delete`	Delete vectors by ID or query	Write
`supavec_describe_index`	Get index schema information	Discovery

Detailed Usage

supavec_upsert▶

Insert new vectors or update existing ones in the Supavec index, optionally with associated metadata.

use_mcp_tool({
  server_name: "supavec",
  tool_name: "supavec_upsert",
  arguments: {
    vectors: [
      {
        id: "doc1",
        values: [0.1, 0.2, 0.3, ...],
        metadata: { "author": "John Doe", "year": 2023 }
      },
      {
        id: "doc2",
        values: [0.4, 0.5, 0.6, ...],
        metadata: { "category": "AI", "status": "published" }
      }
    ]
  }
});

Returns the count of upserted vectors.

supavec_query▶

Perform a similarity search against the Supavec index using a query vector, with options for top-k results and metadata filtering.

use_mcp_tool({
  server_name: "supavec",
  tool_name: "supavec_query",
  arguments: {
    query_vector: [0.1, 0.1, 0.1, ...],
    top_k: 5,
    filter: { "year": { "$eq": 2023 } }
  }
});

Returns a list of matching vectors and their similarity scores.

supavec_delete▶

Delete vectors from the Supavec index by their IDs or by specifying a metadata filter.

// Delete a single vector by ID
use_mcp_tool({
  server_name: "supavec",
  tool_name: "supavec_delete",
  arguments: {
    ids: ["doc1"]
  }
});

// Delete multiple vectors by metadata filter
use_mcp_tool({
  server_name: "supavec",
  tool_name: "supavec_delete",
  arguments: {
    filter: { "status": { "$eq": "draft" } }
  }
});

Returns the number of deleted vectors.

supavec_describe_index▶

Retrieve information about the Supavec index, including its dimensions, vector count, and configured metadata fields.

use_mcp_tool({
  server_name: "supavec",
  tool_name: "supavec_describe_index",
  arguments: {}
});

Returns a JSON object with index details.

Installation

{
  "mcpServers": {
    "supavec": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "mcp/supavec",
        "supavec://host.docker.internal:5000?api_key=YOUR_API_KEY"
      ]
    }
  }
}

Docker Networking:

Use host.docker.internal instead of localhost when connecting from Docker to Supavec on your host machine. Replace YOUR_API_KEY with your actual Supavec API key.

Common Use Cases

1. Retrieval Augmented Generation (RAG)

Enhance LLM responses by retrieving relevant context from your vector database:

// Query Supavec for relevant documents based on a user's question
use_mcp_tool({
  server_name: "supavec",
  tool_name: "supavec_query",
  arguments: {
    query_vector: [0.7, 0.2, 0.9, ...], // Embedding of user's question
    top_k: 3,
    filter: { "document_type": { "$eq": "knowledge_base" } }
  }
});

2. Semantic Search

Implement powerful search functionalities that understand the meaning behind queries:

// Search for products semantically similar to a user's input
use_mcp_tool({
  server_name: "supavec",
  tool_name: "supavec_query",
  arguments: {
    query_vector: [0.3, 0.8, 0.1, ...], // Embedding of product description
    top_k: 10,
    filter: { "availability": { "$eq": "in_stock" } }
  }
});

3. Recommendation Systems

Suggest items or content based on user preferences or item similarities:

// Recommend articles similar to what the user has read
use_mcp_tool({
  server_name: "supavec",
  tool_name: "supavec_query",
  arguments: {
    query_vector: [0.6, 0.4, 0.5, ...], // Embedding of user's read history
    top_k: 5,
    filter: { "category": { "$ne": "read" } }
  }
});

4. Anomaly Detection

Identify unusual data points by comparing their vector embeddings to a baseline:

// Detect anomalous user behavior based on activity embeddings
use_mcp_tool({
  server_name: "supavec",
  tool_name: "supavec_query",
  arguments: {
    query_vector: [0.9, 0.1, 0.2, ...], // Embedding of current user activity
    top_k: 1,
    filter: { "user_status": { "$eq": "active" } }
  }
});

Connection String Format

The Supavec MCP server accepts connection URLs in the following format:

Standard: supavec://host:port?api_key=YOUR_API_KEY
With Namespace: supavec://host:port/namespace?api_key=YOUR_API_KEY
TLS/SSL: supavecs://host:port?api_key=YOUR_API_KEY

Supavec MCP Server: Vector Database for AI Applications