medical-billing
Open-source MCP server providing medical billing knowledge (ICD-10, CPT, denial codes) for AI assistants
claude mcp add --transport stdio kustode-ce-medical-billing-mcp python -m medical_billing_mcp
How to use
The Medical Billing MCP provides a knowledge layer for AI assistants to answer questions about medical billing codes, denials, and payer rules without handling live payer connectivity. It exposes a set of tools that you can invoke through the MCP protocol, including: lookup_icd10 (diagnosis code lookups), lookup_cpt (procedure code lookups), lookup_modifier (modifier usage guidance), lookup_denial (denial code explanations and resolution steps), lookup_payer (payer-specific rules such as timely filing), and lookup_bundling (whether codes are bundled). You can run this server locally or via Docker and connect an AI assistant like Claude or another MCP-enabled agent to access the knowledge base and respond with guidance, resolution steps, and best practices. The server is designed to be used as a knowledge layer; you provide connectivity to payer data (e.g., Stedi, Availity, Change Healthcare, etc.) on your side.
To use it, start the MCP server and configure your agent to point at the medical-billing MCP server. For Claude Desktop users, you can add the server to claude_desktop_config.json under mcpServers with the command python -m medical_billing_mcp. Once connected, ask questions such as, “What does ICD-10 code E11.9 mean?”, “What is CPT code 99214 and what documentation do I need?”, “When should I use modifier 25?”, or “What are the resolution steps for denial CO-50?”. The agent will route queries to the appropriate tool and respond with concise explanations and actionable steps.
How to install
Prerequisites:\n- Python 3.8+ (or a compatible recent Python environment)\n- Git\n\nOption A: Local install (recommended for development)\n1) Clone the repository:\nbash\ngit clone https://github.com/Kustode-ce/medical-billing-mcp.git\ncd medical-billing-mcp\n\n2) Install in editable mode:\nbash\npip install -e .\n\n3) Run the MCP server:\nbash\npython -m medical_billing_mcp\n\n\nOption B: Docker (Docker Compose)\n1) Clone the repository and navigate to it:\nbash\ngit clone https://github.com/Kustode-ce/medical-billing-mcp.git\ncd medical-billing-mcp\n\n2) Run the server with Docker Compose:\nbash\ndocker compose up -d\n\n3) Test the server (in container):\nbash\ndocker compose exec mcp python -m medical_billing_mcp --test\n\n\nNote: If you plan to run via Docker, ensure docker-compose.yml is configured in the docker/ directory as shown in the repository.\n
Additional notes
Tips and considerations:\n- The MCP server uses local JSON data sources (data/icd10.json, data/cpt.json, data/modifiers.json, data/denials.json, data/payers.json, data/bundling.json). You can contribute or update these files to improve accuracy.\n- The server is intentionally a knowledge layer; real-time payer connectivity and submission actions are performed by your integrated systems (e.g., Stedi, Availity, Change Healthcare).\n- For Claude Desktop users, configure the server in claude_desktop_config.json under mcpServers with the following example: {"medical-billing": {"command": "python", "args": ["-m", "medical_billing_mcp"]}}.\n- You can test basic functionality locally by running python -m medical_billing_mcp --test (if supported by the repo) or by exercising the tooling via your MCP client.\n- When deploying, consider exposing logging levels and enabling verbose mode during initial integration to diagnose tool routing issues.\n
Related MCP Servers
mysql_mcp_server
A Model Context Protocol (MCP) server that enables secure interaction with MySQL databases
Gitingest
mcp server for gitingest
fhir
FHIR MCP Server – helping you expose any FHIR Server or API as a MCP Server.
skill-to
Convert AI Skills (Claude Skills format) to MCP server resources - Part of BioContextAI
scraper
Context-optimized MCP server for web scraping. Reduces LLM token usage by 70-90% through server-side CSS filtering and HTML-to-markdown conversion.
ros2_medkit_mcp
MCP server for ros2_medkit. Bridge LLM agents to the SOVD REST API for ROS 2 diagnostics and remote operations.