API Docs Query Server
Published on July 22nd, 2025Overview
The API Docs Query Server provides AI assistants with comprehensive access to your organization’s API documentation and specifications stored in your Backstage catalog.
Server Endpoint: https://api.roadie.so/api/mcp/v1/api-docs-query
Capabilities
- API Discovery: Search for APIs using natural language queries
- Specification Retrieval: Get complete OpenAPI, GraphQL, or AsyncAPI specifications
- Intelligent Search: Find APIs by domain, technology, or business context
- Metadata Access: Retrieve API descriptions, ownership, and categorization
Available Tools
Find API Specs
Search for available API specifications using a query string that supports partial matching across API names, descriptions, and metadata.
Parameters:
queryString
(string): Search term for finding API specs
Example Usage:
{
"queryString": "payment"
}
Return Schema:
{
results: {
type: string,
document: {
kind: string,
text: string,
type: string,
owner: string,
title: string,
keywords: string,
location: string,
lifecycle: string,
namespace: string,
componentType: string
}
}[]
}
This will return all APIs related to payments, including services like “payment-gateway”, “payment-processor”, or “billing-api”.
Retrieve API Spec
Get the complete specification for a specific API, including full OpenAPI/Swagger definitions, schemas, and endpoint documentation.
Parameters:
name
(string): API namenamespace
(string, optional): API namespace (defaults to “default”)
Example Usage:
{
"name": "user-service-api",
"namespace": "backend"
}
Return Schema:
{
entityRef: string,
spec: string
}
Required Permissions
- Catalog entity read (*) - Access to catalog entities and API specifications
Common Use Cases
API Integration Planning
- Ask your AI assistant: “What payment APIs are available?”
- Get detailed endpoint information for integration planning
- Compare different APIs to choose the best fit
Documentation Exploration
- “Show me the schema for creating a user account”
- “What authentication does the order API require?”
- “List all the endpoints in the inventory service”
Development Assistance
- Generate client code from API specifications
- Create automated tests based on API schemas
- Validate API requests and responses
Examples
Example AI Conversation
User: “I need to integrate with our user management system”
AI Response using MCP:
- Searches for user-related APIs using
find-api-specs
- Retrieves specifications for relevant APIs
- Explains available endpoints, authentication, and schemas
- Provides integration guidance and code examples
Practical Usage
// Finding payment-related APIs
{
"tool": "find-api-specs",
"arguments": {
"queryString": "payment processing"
}
}
// Getting complete specification
{
"tool": "retrieve-api-spec",
"arguments": {
"name": "payment-gateway-api",
"namespace": "payments"
}
}
Best Practices
- Start with broad search terms and refine based on results
- Use domain-specific language (e.g., “payment”, “authentication”, “notification”)
- Check multiple namespaces if APIs aren’t found in default
- Review complete specifications before starting integration