Overview

The Morningstar MCP Server implements a secure, scalable API layer for accessing Morningstar research, editorial content, and datapoints. It exposes search and retrieval endpoints that centralize access to investment research, structured datapoints (facts and metrics), and editorial metadata — enabling product teams to integrate Morningstar content into web apps, analytics pipelines, and downstream services.

Designed as a RESTful service, the MCP server is built to be deployed in modern infrastructure (containers, orchestration) and integrates common production concerns: authentication, pagination, filtering, rate limiting, and caching. Developers can use the server to deliver contextual content in search results, populate UI detail pages, and feed analytics systems with consistent datapoint shapes.

Features

RESTful API for search, document retrieval, and datapoints
Secure access using API keys / Bearer tokens
Full-text and faceted search across research and editorial content
Structured datapoint APIs for financial facts and metrics
Pagination, sorting, and filtering support
Response headers for caching and rate-limit status
Configurable via environment variables and suitable for containerized deployment
Logging and basic observability hooks (metrics-compatible)
Scales horizontally; stateless service design for load balancers

Installation / Configuration

Clone the repository and run locally or in a container. Replace environment values with your configuration.

Clone and build locally:

git clone https://github.com/Morningstar/morningstar-mcp-server.git
cd morningstar-mcp-server
# Build a Docker image (example)
docker build -t morningstar-mcp-server .

Run with Docker:

docker run -d \
  -p 8080:8080 \
  -e MCP_PORT=8080 \
  -e MCP_DB_URL="postgres://user:pass@db:5432/mcp" \
  -e MCP_AUTH_SECRET="your-api-secret" \
  -e MCP_CACHE_TTL=300 \
  --name mcp-server \
  morningstar-mcp-server

Example docker-compose (development):

version: "3.8"
services:
  mcp-server:
    build: .
    ports:
      - "8080:8080"
    environment:
      MCP_PORT: 8080
      MCP_DB_URL: "postgres://user:pass@db:5432/mcp"
      MCP_AUTH_SECRET: "change-me"
      MCP_CACHE_TTL: 300
    depends_on:
      - db

  db:
    image: postgres:13
    environment:
      POSTGRES_USER: user
      POSTGRES_PASSWORD: pass
      POSTGRES_DB: mcp
    volumes:
      - db-data:/var/lib/postgresql/data

volumes:
  db-data:

Common environment variables

MCP_PORT — HTTP port (default 8080)
MCP_DB_URL — database connection string for the content store
MCP_AUTH_SECRET — API key or secret used to validate tokens
MCP_RATE_LIMIT — requests per minute per API key
MCP_CACHE_TTL — default seconds to cache responses

Available Resources

The server exposes versioned REST endpoints under /v1. Typical resource groups:

/v1/search — full-text and filtered search across research and editorial content
/v1/documents/{id} — retrieve a full research or editorial document by ID
/v1/datapoints — query structured datapoints (metrics, facts) by identifiers, namespaces and date ranges
/v1/batch — batch retrieval for multiple IDs (where supported)
/v1/health — service health and readiness probes

API documentation and schema files (OpenAPI/Swagger) are available in the GitHub repository for reference and to generate client SDKs. See: https://github.com/Morningstar/morningstar-mcp-server

Example Requests

Search (keyword + filter)

curl -H "Authorization: Bearer $MCP_TOKEN" \
  "https://mcp.example.com/v1/search?q=income+funds&filter[type]=research&page=1&per_page=20"

Get document

curl -H "Authorization: Bearer $MCP_TOKEN" \
  "https://mcp.example.com/v1/documents/12345"

Query datapoints (by identifier and date range)

curl -H "Authorization: Bearer $MCP_TOKEN" \
  "https://mcp.example.com/v1/datapoints?identifier=MSFT&metric=earnings_per_share&from=2020-01-01&to=2023-12-31"

Example search response (abbreviated)

{
  "total": 234,
  "page": 1,
  "per_page": 20,
  "results": [
    {
      "id": "12345",
      "title": "Quarterly Earnings Review",
      "type": "research",
      "snippet": "Earnings beat expectations due to...",
      "published_at": "2024-01-15T12:00:00Z"
    }
  ]
}

Use Cases

Embed research search into a web portal: Use the /v1/search endpoint to show ranked results with snippets and facets (by asset class, region, analyst).
Detail pages for instruments: On navigation to a security page, fetch the latest editorial report via /v1/documents/{id} and load structured datapoints (financial metrics) from /v1/datapoints for charting and analytics.
Data export for analytics: Periodically call /v1/datapoints with date ranges to populate a data warehouse or feed time-series models.
Workflow automation & compliance: Automate content checks by pulling editorial metadata and timestamps to audit publication pipelines.

Best Practices

Authenticate all requests and rotate API keys periodically.
Use pagination to limit response sizes; avoid requesting large per_page values.
Respect rate-limit headers; implement exponential backoff on 429 responses.
Cache immutable responses (document snapshots) using Cache-Control headers and respect TTL configuration.
Run the server behind a load balancer and use health/readiness endpoints for orchestration.

For full API reference and schema downloads, see the project repository: https://github.com/Morningstar/morningstar-mcp-server

Morningstar MCP Server: Research, Editorial, Datapoints