Build with AI

Stedi provides AI-powered tools and resources to help you build eligibility and claims processing workflows faster.

Build agents that work with eligibility data: Install the Model Context Protocol (MCP) server
Code with AI assistants: Give them AI-friendly versions of our developer guides and API specifications, and use test workflows to develop without real patient data

Get started

Try building with Stedi's AI-friendly resources by following our step-by-step tutorial. You'll build a functional insurance verification app in under 30 minutes using Stedi's JSON Eligibility API and a coding agent like Claude Code or Cursor.

The tutorial walks through the complete process: setting up your Stedi account, generating test API keys, building a Next.js backend with AI assistance, and creating a web UI. You'll use the same AI-friendly documentation and test workflows described on this page.

Build an insurance verification app with a coding agent

Test workflows

Stedi provides test workflows that AI coding agents can use to build and test integrations without using real patient data or contacting actual payers.

Get a test API key

Test API keys allow you to send mock requests and receive realistic responses without transmitting PHI or PII.

To create a test API key:

Log into your Stedi account.
Click your account name at the top right.
Select API Keys.
Click Generate new API Key.
Enter a name with a test prefix and choose Test as the key type.
Click Generate and copy your key.

Learn more about API authentication.

Test eligibility checks

Use your test API key to send mock eligibility requests that return realistic responses including coverage indicators, copays, and deductibles. Mock requests work with Stedi's JSON, Raw X12, and SOAP endpoints. Test keys are free on the Basic plan.

Here's an example test request your AI assistant can use:

curl --request POST \
  --url 'https://healthcare.us.stedi.com/2024-04-01/change/medicalnetwork/eligibility/v3' \
  --header 'Authorization: Key {test_api_key}' \
  --header 'Content-Type: application/json' \
  --data '{
    "tradingPartnerServiceId": "60054",
    "provider": {
        "organizationName": "Provider Name",
        "npi": "1999999984"
    },
    "subscriber": {
        "firstName": "John",
        "lastName": "Doe",
        "memberId": "AETNA9wcSu"
    },
    "dependents": [
        {
            "firstName": "Jordan",
            "lastName": "Doe",
            "dateOfBirth": "20010714"
        }
    ],
    "encounter": {
        "serviceTypeCodes": ["30"]
    }
}'

Test claims

Use production API keys with the test usage indicator ("usageIndicator": "T") to submit test claims. Claims aren't forwarded to payers, and you receive test 277CA acknowledgments. Submit to the Stedi Test Payer (ID: STEDI) to generate test ERAs within minutes.

Visit Test claims workflow for complete instructions.

Data security and compliance

When using AI tools with healthcare data, follow your organization's data handling policies to ensure HIPAA compliance. Most organizations require a Business Associate Agreement (BAA) with third-party AI tools before using them with production healthcare data.

Stedi's test workflows let you develop and test integrations without transmitting PHI or PII.

Model Context Protocol server

The Model Context Protocol (MCP) server provides tools that AI agents can use to perform and troubleshoot eligibility checks through Stedi's APIs. It includes built-in logic for finding the correct payer, handling errors, and implementing retry strategies, so you can focus on building your agent's core functionality.

The MCP server is ideal for:

Voice agents that need to verify benefits in real time
Revenue Cycle Management (RCM) workflow agents that validate coverage before scheduling appointments
Development teams using MCP clients to test integrations and troubleshoot issues

The MCP server supports both API key and OAuth 2.x authentication, and works with popular AI platforms like Claude Code.

AI-friendly documentation

Stedi's documentation is available in AI-friendly formats that you can provide to tools like ChatGPT, Claude, Cursor, and other AI assistants.

For best results, provide both OpenAPI specifications and llms-full.txt for comprehensive developer guides.

OpenAPI specifications

github.com/Stedi/openApi

OpenAPI specifications provide machine-readable API definitions including endpoints, parameters, request/response schemas, and real-world examples.

When to use: Provide these to your AI assistant when you need it to help you generate API calls, understand API structure, or validate request and response formats.

llms.txt

/llms.txt

A simple list of available documentation pages with titles and descriptions.

When to use: Use this to give your AI assistant a quick overview of available pages. Many AI assistants check for this file automatically.

llms-full.txt

/llms-full.txt

The complete markdown content of all documentation pages in a single file. This includes guides, how-tos, troubleshooting, and conceptual explanations.

When to use: Provide this to your AI assistant when you need comprehensive context about how to build eligibility and claims processing workflows with Stedi.

Individual markdown pages

Individual documentation pages are available in markdown format at /llms.mdx/[page-url]. For example, /llms.mdx/healthcare/send-eligibility-checks. You can also copy them from the UI by clicking the Copy for LLM menu at the top of each documentation page.

When to use: Use individual pages when you have token limits that prevent loading llms-full.txt, or when you only need context about specific features.

How to use these resources

The specific method depends on which AI tool you're using.

You may want to add references to your agent's configuration - for example, a Cursor rule or Claude skill file. This ensures the context is applied automatically across sessions.

You have access to Stedi's healthcare API documentation and specifications:
- Full documentation: https://www.stedi.com/docs/llms-full.txt
- OpenAPI specs: https://github.com/Stedi/openApi

When helping with Stedi integrations, consult these resources first.

You can also provide these resources as context when starting a conversation with tools like ChatGPT, Claude, or Google Gemini.

I'm working with Stedi's healthcare APIs. Please reference:
- Full documentation: https://www.stedi.com/docs/llms-full.txt
- OpenAPI specs: https://github.com/Stedi/openApi

Help me [describe your task].

Example prompts

Once your AI assistant has access to the documentation, you can ask it to help with tasks like:

"Help me implement eligibility checks for a new RCM system. Show me how to construct the JSON request and handle the response."
"Generate TypeScript code to submit professional claims using Stedi's API, including error handling and retry logic."
"What's the correct Claim Frequency Code for resubmitting an institutional claim that was already adjudicated?"

Always review and test AI-generated code before using it in production. While AI assistants can help you build integrations faster, you should verify that the generated code meets your specific requirements and follows your organization's standards.

Build with AI

On this page