Tax Document

mcp server sample file입니다.

python

0

GitHub Stars

python

Language

7 months ago

First Indexed

3 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "cheolminbae-mcp-server-sample": {
      "command": "python",
      "args": [
        "mcp_server.py"
      ],
      "env": {
        "PDF_PATH": "2025_tax.pdf",
        "CHROMA_HOST": "localhost",
        "CHROMA_PORT": "8000"
      }
    }
  }
}

You run an MCP server that indexes a tax law PDF and makes its content searchable through natural language queries. It uses a vector store to accelerate retrieval and exposes an MCP API so clients can search, retrieve document info, and list indexed collections with stable, predictable commands.

How to use

Interact with the MCP server from an MCP client to perform document searches, fetch information, and manage stored collections. You can search the document content using natural language queries, retrieve metadata about the indexed document, and list all stored collections. The server responds with relevant passages and structured results, enabling you to build search experiences, Q&A flows, or knowledge tools around the 2025 tax law content.

How to install

# Prerequisites
# - Python 3.8+
# - Docker and Docker Compose (for the vector store)
# - git (optional, if cloning)

# 1. Install Python dependencies
pip install -r requirements.txt

# 2. Start ChromaDB (vector store) locally
docker-compose up -d

# 3. Index the PDF document into the vector store
python index_document.py

# 4. Run the MCP server
python mcp_server.py

Additional setup and notes

Environment variables that influence the server and its vector store include: CHROMA_HOST, CHROMA_PORT, and PDF_PATH. Set these to configure where the vector store runs and which PDF is indexed. Common defaults are localhost for CHROMA_HOST, 8000 for CHROMA_PORT, and a path like 2025_tax.pdf for PDF_PATH.

Available tools

search_document

Searches the indexed document content using natural language queries and returns relevant passages or results.

get_document_info

Retrieves metadata and basic information about the indexed document.

list_collections

Lists all stored vector collections available in the MCP server.