Web Search and Semantic Similarity MCP server for AI agents

Parquet MCP Server is a specialized Model Control Protocol server designed to enhance Claude Desktop with powerful search capabilities. It offers web search functionality with result scraping and similarity searching to extract relevant information from previous searches, making it particularly useful for applications requiring web search integration.

Installation Options

Via Smithery (Recommended)

The easiest way to install Parquet MCP Server for Claude Desktop is through Smithery:

npx -y @smithery/cli install @DeepSpringAI/parquet_mcp_server --client claude

Manual Installation

If you prefer manual installation:

Clone the repository and navigate to it:

git clone https://github.com/path/to/parquet_mcp_server.git
cd parquet_mcp_server

Create and activate a virtual environment:

uv venv
# On Windows
.venv\Scripts\activate
# On macOS/Linux
source .venv/bin/activate

Install the package:

uv pip install -e .

Configuration

Environment Setup

Create a .env file with the following variables:

EMBEDDING_URL=http://sample-url.com/api/embed
OLLAMA_URL=http://sample-url.com/
EMBEDDING_MODEL=sample-model
SEARCHAPI_API_KEY=your_searchapi_api_key
FIRECRAWL_API_KEY=your_firecrawl_api_key
VOYAGE_API_KEY=your_voyage_api_key
AZURE_OPENAI_ENDPOINT=http://sample-url.com/azure_openai
AZURE_OPENAI_API_KEY=your_azure_openai_api_key

Claude Desktop Integration

Add the server to your Claude Desktop configuration file (claude_desktop_config.json):

{
  "mcpServers": {
    "parquet-mcp-server": {
      "command": "uv",
      "args": [
        "--directory",
        "/home/${USER}/workspace/parquet_mcp_server/src/parquet_mcp_server",
        "run",
        "main.py"
      ]
    }
  }
}

Using the MCP Server

Available Tools

The server provides two primary tools:

Search Web: Performs web searches and scrapes results
- Required: queries (list of search queries)
- Optional: page_number (defaults to 1)
Extract Info from Search: Finds and extracts relevant information from previous searches
- Required: queries (list of search queries to match against)

Example Prompts

For web searching:

"Please perform a web search for 'macbook' and 'laptop' and scrape the results from page 1"

For extracting information:

"Please extract relevant information from the previous searches for 'macbook'"

Programmatic Usage

You can also use the client directly in your Python code:

from parquet_mcp_server.client import (
    perform_search_and_scrape,
    find_similar_chunks
)

# Perform a web search
perform_search_and_scrape(["macbook", "laptop"], page_number=1)

# Extract information from the search results
find_similar_chunks(["macbook"])

Troubleshooting

If you encounter issues:

SSL Verification Errors: Check SSL settings in your .env file
Embedding Generation Problems: Verify that:
- The Ollama server is running and accessible
- The specified model is available on your Ollama server
- The text column exists in your input Parquet file
DuckDB Conversion Issues: Ensure:
- The input Parquet file exists and is readable
- You have write permissions for the output directory
- The Parquet file isn't corrupted
PostgreSQL Conversion Problems: Check:
- PostgreSQL connection settings in your .env file
- The PostgreSQL server is running and accessible
- You have permissions to create/modify tables
- The pgvector extension is installed in your database

Database Configuration

PostgreSQL Table Setup

For PostgreSQL integration, create the necessary table:

CREATE TABLE web_search (
    id SERIAL PRIMARY KEY,
    text TEXT,
    metadata JSONB,
    embedding VECTOR(1024),
    date TIMESTAMP DEFAULT NOW()
);

Similarity Search Function

For vector similarity searches in PostgreSQL:

CREATE OR REPLACE FUNCTION match_web_search(
  query_embedding vector(1024),
  match_threshold float,
  match_count int
)
RETURNS TABLE (
  id bigint,
  metadata jsonb,
  text TEXT,
  date TIMESTAMP,
  similarity float
)
LANGUAGE plpgsql
AS $$
BEGIN
  RETURN QUERY
  SELECT
    web_search.id,
    web_search.metadata,
    web_search.text,
    web_search.date,
    1 - (web_search.embedding <=> query_embedding) as similarity
  FROM web_search
  WHERE 1 - (web_search.embedding <=> query_embedding) > match_threshold
  ORDER BY web_search.date DESC,
           web_search.embedding <=> query_embedding
  LIMIT match_count;
END;
$$;

How to add this MCP server to Cursor

There are two ways to add an MCP server to Cursor. The most common way is to add the server globally in the ~/.cursor/mcp.json file so that it is available in all of your projects.

If you only need the server in a single project, you can add it to the project instead by creating or adding it to the .cursor/mcp.json file.

Adding an MCP server to Cursor globally

To add a global MCP server go to Cursor Settings > MCP and click "Add new global MCP server".

When you click that button the ~/.cursor/mcp.json file will be opened and you can add your server like this:

{
    "mcpServers": {
        "cursor-rules-mcp": {
            "command": "npx",
            "args": [
                "-y",
                "cursor-rules-mcp"
            ]
        }
    }
}

Adding an MCP server to a project

To add an MCP server to a project you can create a new .cursor/mcp.json file or add it to the existing one. This will look exactly the same as the global MCP server example above.

How to use the MCP server

Once the server is installed, you might need to head back to Settings > MCP and click the refresh button.

The Cursor agent will then be able to see the available tools the added MCP server has available and will call them when it needs to.

You can also explictly ask the agent to use the tool by mentioning the tool name and describing what the function does.