Model Context Protocol (MCP)

Connect AI agents to the RCI knowledge layer using the Model Context Protocol

Our MCP server exposes the RCI knowledge layer as a set of tools that AI agents can use to resolve billing context, calculate payments, research codes, and resolve denials.

When building agents that work with healthcare billing data, we recommend using our MCP server. It excels at knowledge-driven tasks where your agent needs billing context in real time. For example:

• An RCM workflow agent can use the MCP server to resolve the complete billing context before constructing a claim or writing an appeal.
• A coding assistant can use the MCP server to research CPT/HCPCS codes, validate medical necessity, and check documentation requirements.
• A denial management agent can use the MCP server to look up CARC/RARC codes, retrieve payer appeal rules, and generate resolution strategies.

Why use the MCP server?

Resolving healthcare billing knowledge correctly requires navigating six layers of context — from geographic payment adjustments (GPCI) to payer-specific filing deadlines to service-level RVU calculations. Getting any of these wrong leads to underpayment, denials, or compliance violations.

The MCP server handles this complexity so your agent doesn’t have to. It comes with:

• Full L1-L6 resolution from a single CCN and CPT code
• Payment calculation using the MPFS methodology with GPCI adjustments
• Payer intelligence including timely filing limits, appeal levels, and prior auth rules
• Code research via the async code-research agent
• Denial resolution with CARC/RARC lookup and appeal strategy generation
• Cited policy, 340B, and prior auth — same tools as REST (lookup_policy, resolve_340b_restrictions, check_pa_required)

Features

The MCP server allows agents to:

• Resolve billing context: The resolve_knowledge tool chains all six layers, returning GPCI values, payment systems, billing forms, payer rules, and payment estimates.
• Calculate expected payment: The calculate_payment tool computes MPFS-based payment using GPCI-adjusted RVUs and the current conversion factor.
• Get billing guidelines: The get_billing_guide tool returns the correct billing form, TOB codes, modifier requirements, and any exceptions.
• Research billing codes: The research_code tool submits a query to the AI code research agent and returns structured results.
• Resolve denials: The resolve_denial tool analyzes CARC/RARC codes and generates appeal strategies with payer-specific deadlines.
• Search payer rules: The search_payer tool retrieves payer-specific rules including timely filing, appeal levels, and prior auth requirements.

Limitations

The MCP server is designed for agent workflows that need billing knowledge for individual encounters. It is not intended for:

• Batch processing — Use the REST API directly for bulk resolution
• Claims submission — RCI provides knowledge, not clearinghouse connectivity. Use Stedi or your clearinghouse for claims submission.
• Real-time eligibility — RCI resolves billing rules, not patient-specific eligibility. Use Stedi for eligibility checks.

Data security

The MCP server uses TLS for all connections and authenticates with the same X-API-Key as REST.

How the MCP server works

Your AI agent connects to the server using an MCP client. Once connected, the server gives your agent access to the available tools and prompts.

┌─────────────────┐     MCP Protocol      ┌─────────────────────┐
│   Your Agent    │◄─────────────────────►│   RCI MCP Server    │
│ (Claude, Cursor,│    tools + prompts     │                     │
│  custom agent)  │                        │  resolve_knowledge  │
└─────────────────┘                        │  calculate_payment  │
                                           │  get_billing_guide  │
                                           │  research_code      │
                                           │  resolve_denial     │
                                           │  search_payer       │
                                           └──────────┬──────────┘
                                                      │
                                           ┌──────────▼──────────┐
                                           │   RCI Knowledge     │
                                           │   Layer API         │
                                           │   (L1-L6 Graph)     │
                                           └─────────────────────┘

Tools

Tool Name	Description
resolve_knowledge	Full L1-L6 resolution for a facility and optional service. Returns GPCI, payment system, billing form, payer rules, RVUs, and payment estimate.
calculate_payment	MPFS-based payment calculation with GPCI adjustments. Returns estimated payment and full RVU component breakdown.
get_billing_guide	Billing form, TOB codes, modifier requirements, and consolidated billing rules for a facility and service.
research_code	AI-powered code research for CPT, HCPCS, and ICD-10 codes. Returns coding guidance with documentation requirements.
resolve_denial	Analyzes CARC/RARC denial codes and generates payer-specific appeal strategies with deadlines.
search_payer	Searches RCI’s payer database for rules, timely filing limits, appeal levels, and prior auth requirements.

Prompts

The MCP server includes prompts to help your agent with common billing workflows:

• How to interpret a knowledge resolution response
• When to use facility vs. non-facility rates
• How to handle unknown CPT codes
• Best practices for denial appeal construction

Pricing

The MCP server is available on all paid RCI plans. There is no charge for using the MCP server itself — you are only charged for the underlying API requests.