Skip to main content

​
MCP Server — Trust Query Tool

​
7.1 What Is MCP

MCP (Model Context Protocol) is a standard protocol that connects AI applications with external tools and data sources. Initiated by Anthropic, it is now governed by the Linux Foundation. In plain terms: MCP is the “USB-C port” for AI. Just as USB-C lets your phone connect to chargers, monitors, and flash drives, MCP lets AI connect to a wide variety of external tools and data sources. Through MCP, an AI agent can:
  • Query databases
  • Call APIs
  • Read files
  • Execute actions
For the full MCP technical specification, see Book 5: MCP Protocol Deep Dive. This chapter focuses solely on the OTR trust query MCP Server implementation.

​
7.2 OTR MCP Server

OTR provides an MCP Server that enables any MCP-compatible AI application (Claude, ChatGPT, etc.) to invoke trust query capabilities directly.

​
Available Tools

Tool NameFunctionInputOutput
otr_verifyQuery a domain’s trust scoredomain: stringComplete OTR scoring data
otr_compareCompare multiple domainsdomains: string[]Side-by-side score comparison
otr_signalsView signal detailsdomain: string, dimension: stringDetailed signals for the specified dimension

​
MCP Server Connection

{
  "mcpServers": {
    "otr": {
      "url": "https://orbexa.io/mcp"
    }
  }
}

​
7.3 Using OTR in Claude

If you use Claude Desktop or Claude Code, you can add the OTR Server in your MCP settings: Configuration file (claude_desktop_config.json): 
{
  "mcpServers": {
    "otr-trust": {
      "url": "https://orbexa.io/mcp"
    }
  }
}
Once configured, you can simply say in a conversation: 
Look up the trust score for amazon.com and give me a detailed analysis of each dimension.
Claude will automatically invoke the otr_verify tool on the OTR MCP Server, retrieve the data, and present its analysis.

​
7.4 Mintlify Auto-MCP

Mintlify (the documentation platform powering this knowledge base) automatically generates an MCP Server for all hosted documentation. This means AI agents can query the entire contents of learn.orbexa.io via the MCP protocol. This is a defining feature of AI-native documentation platforms: the documentation itself becomes a data source that AI agents can programmatically access.

​
7.5 Building Your Own MCP Server

If you want to build an MCP Server for your own trust assessment system, the basic architecture is as follows: 
AI Application (Claude / ChatGPT)
    | MCP Protocol
MCP Server (your service)
    | HTTP
OTR API or your custom assessment API
    |
Returns scoring data
The core of an MCP Server is a set of Tools, each defined by:
  • A name and description
  • An input parameter JSON Schema
  • Execution logic (calling an API to fetch data)
For a comprehensive MCP development guide, see Book 5: MCP Protocol Deep Dive.

​
7.6 MCP vs. REST API: When to Use Which

ScenarioRecommendedRationale
Querying within an AI conversationMCPAI-native experience; no manual API calls needed
Programmatic integrationREST APIStandard HTTP calls; supported by any language
Batch queriesREST APIMore flexible control over request flow
Autonomous AI agent decisionsMCPAgent invokes queries directly without human intervention
Dashboards and reportsREST APIScheduled data pulls for display

​
7.7 Core Principle: Autonomous Agent Queries

The value of an MCP Server is not simply “one more way to make API calls.” It lies in enabling AI agents to autonomously decide when to query trust data. Traditional workflow: The user manually opens a website, types in a domain, and reviews the results. MCP workflow: The user tells the AI, “Find me a reliable outdoor gear supplier,” and the AI agent automatically:
  1. Searches for candidate merchants
  2. Invokes OTR via MCP to query each merchant’s trust score
  3. Ranks candidates by a composite of score, price, and reviews
  4. Explains its recommendation rationale to the user
Throughout this process, the user never needs to know that OTR exists. Trust assessment operates as an automatic data input within the AI’s decision-making pipeline.
Next chapter: OpenAPI Specification — The OpenAPI 3.0 specification for the OTR API