MCP Fact-Check

Validates MCP content against official specifications using semantic search and AI-powered validation.

GitHub Stars

Language

4 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "carlisia-mcp-factcheck": {
      "command": "/path/to/bin/mcp-factcheck-server",
      "args": [
        "--data-dir",
        "/path/to/data/embeddings",
        "--telemetry",
        "--otlp-endpoint",
        "http://localhost:6006"
      ],
      "env": {
        "GITHUB_TOKEN": "YOUR_GITHUB_TOKEN",
        "OPENAI_API_KEY": "YOUR_API_KEY"
      }
    }
  }
}

You use the MCP Fact-Check MCP Server to validate content against official MCP specifications, helping you ensure technical accuracy and avoid misinformation. It leverages semantic search and AI-powered validation to assess MCP-related claims, supports multiple specification versions, and can break down complex statements for precise checks.

How to use

You connect your MCP client to the server and submit MCP-related content you want to validate. The server analyzes multi-claim content, decomposes compound statements, and returns validation results with confidence scores. It also offers quick fact checks for single claims and can search specification sections to speed up your review.

How to install

Install the MCP Fact-Check server through your MCP client’s server marketplace. Find the MCP Fact-Check server named accordingly, install it, and provide your OpenAI API key when prompted. After installation, the server is automatically configured and ready to use.

If you are developing or contributing, you can explore the project’s development setup for building and testing locally.

Additional configuration and samples

The server exposes a standard runtime configuration that you can integrate into your MCP client or orchestration layer. Your environment must include an OpenAI API key to enable embedding generation and content validation.

{
  "mcpServers": {
    "mcp-factcheck": {
      "command": "/path/to/bin/mcp-factcheck-server",
      "args": [
        "--data-dir",
        "/path/to/data/embeddings",
        "--telemetry",
        "--otlp-endpoint",
        "http://localhost:6006"
      ],
      "env": {
        "OPENAI_API_KEY": "YOUR_API_KEY"
      }
    }
  }
}

Observability with AI-focused tracing

For enhanced tracing and visualization, you can set up an AI-focused observability UI named Arize Phoenix. This helps you view the validation pipeline timeline, embedding performance, and per-chunk validation confidence.

# Install Phoenix
pipx install arize-phoenix

# Start Phoenix server
phoenix serve

Notes on deployment and versions

The server supports multiple MCP specification versions, including drafts and historical releases, and can guide content migration between versions. You can query available versions and migrate content as needed.

Available tools

check_mcp_claim

Validates multi-claim MCP content, decomposes compound statements, and returns corrected content with confidence scores.

check_mcp_quick_fact

Performs quick fact checks for single MCP claims with concise verdict and explanation.

search_spec

Searches MCP specifications by semantic similarity across versions.

list_spec_versions

Lists available MCP specification versions with dates and current status.

migrate-mcp-content

Guides content migration between MCP specification versions with step-by-step prompts and preservation of tone.

Built by

VeilStrat

AI signals for GTM teams