API Reference¶

This section provides detailed documentation for all enrichmcp components.

Core Components¶

EnrichMCP ¶

The main application class that manages entities, relationships, and resources.

from enrichmcp import EnrichMCP

app = EnrichMCP(title="My API", description="API for AI agents")

EnrichModel ¶

Base class for all entities, extending Pydantic's BaseModel with relationship support.

from enrichmcp import EnrichModel
from pydantic import Field


@app.entity
class User(EnrichModel):
    id: int = Field(description="User ID")
    name: str = Field(description="Username")

Relationship ¶

Defines connections between entities with automatic resolver registration.

from enrichmcp import Relationship


class User(EnrichModel):
    orders: list["Order"] = Relationship(description="User's orders")

EnrichContext ¶

Placeholder for future context functionality (currently minimal implementation).

Errors ¶

Currently uses standard Python exceptions and Pydantic validation.

Key Concepts¶

Entity Registration¶

Entities must be registered with the app using the @app.entity decorator:

@app.entity
class Product(EnrichModel):
    """Product in our catalog."""

    id: int = Field(description="Product ID")
    name: str = Field(description="Product name")

Relationship Resolution¶

Every relationship needs a resolver function:

@User.orders.resolver
async def get_user_orders(user_id: int) -> list["Order"]:
    """Fetch orders for a user."""
    return fetch_orders_for_user(user_id)

Resource Creation¶

Resources are the entry points for AI agents:

@app.resource
async def list_users() -> list[User]:
    """List all users in the system."""
    return fetch_all_users()

Type Safety¶

enrichmcp enforces strict type checking:

Entity Validation: All fields must have descriptions
Resolver Validation: Return types must match relationship types exactly
Runtime Validation: Pydantic validates all data at runtime

Schema Introspection¶

AI agents can explore your data model using the built-in explore_data_model() resource:

# Automatically available to AI agents
result = await explore_data_model()
# Returns comprehensive schema information

Best Practices¶

Always include descriptions - Every entity, field, and relationship needs a description
Use meaningful names - Clear, descriptive names help AI agents understand your model
Keep resolvers simple - Each resolver should do one thing well
Handle missing data gracefully - Return sensible defaults rather than errors when possible