Using Beltic MCP

The Beltic MCP Server provides a structured interface for connecting AI clients and automation tools to Beltic’s APIs and datasets using the Model Context Protocol (MCP).

It enables Large Language Models (LLMs), assistants, and internal automation systems to securely query and act on live Beltic data — without requiring custom SDKs or manual API integration.

What is Beltic MCP?

Beltic MCP is Beltic’s official implementation of the Model Context Protocol.
It exposes a schema-aware interface to Beltic’s core products — Identity, Graph, Risk, and Workflows — through a standardized MCP endpoint.

Endpoint:

https://mcp.beltic.com

Through this endpoint, approved clients can:

Query live operational data (e.g., verifications, risk cases, entities)
Access metadata and documentation in structured form
Execute authorized actions (e.g., create workflows, trigger KYC checks)
Receive automatic schema synchronization as Beltic APIs evolve

Architecture Overview

Layer	Description
Context Broker	Translates Beltic API and data structures into standardized MCP contexts for AI and automation clients.
Schema Registry	Maintains synchronized schema definitions for Beltic APIs.
Policy Engine	Enforces role-based access control (RBAC) and ensures scoped permissions for every operation.
Execution Gateway	Executes authorized API actions via signed and auditable requests.
Observability Layer	Logs and monitors MCP activity for audit and performance metrics.

All operations through Beltic MCP are authenticated, audited, and governed by internal access control policies.

Supported Clients

Beltic MCP currently supports integration via Cursor (Alpha).
Support for additional MCP-compatible clients will be added in future releases.

Cursor (Alpha)

You can connect Beltic MCP to Cursor to enable AI-assisted querying of Beltic data and documentation.

Add the configuration to your

.cursor/mcp.json

file:

{
  "mcpServers": {
    "beltic": {
      "url": "https://mcp.beltic.com"
    }
  }
}

After saving, Cursor will automatically detect the MCP server and attempt to connect.

When prompted, authorize access to your Beltic workspace.

Note: Beltic MCP access requires authentication credentials issued by Beltic.

Example Usage

After connecting through Cursor, you can issue natural-language queries that Beltic MCP translates into structured API calls, such as:

“Show all failed verifications this week” → Retrieves filtered Identity API data
“Generate a risk summary for active cases” → Fetches data from Beltic Risk endpoints
“Create a new workflow for user 12345” → Executes a Workflow API action with RBAC enforcement