BigShield MCP Server

An MCP (Model Context Protocol) server that gives AI agents access to BigShield's email validation API. Agents can validate emails, check domain reputation, monitor usage, and submit feedback — all through standard MCP tool calls.

Tools

| Tool | Description | |------|-------------| | validate_email | Validate a single email for spam/fraud risk. Returns risk score, recommendation, fraud decision, and detailed signals. | | validate_emails_batch | Validate up to 100 emails in one request. | | get_validation | Retrieve a previous validation result by ID (useful for polling Tier 2 completion). | | check_domain | Free domain reputation lookup — no API key required. | | get_usage | Check your API usage and remaining quota. | | submit_feedback | Report false positives/negatives to improve accuracy. |

Setup

1. Get an API key

Sign up at bigshield.app and create an API key from your dashboard.

2. Configure your MCP client

Add this to your MCP client config (e.g. claude_desktop_config.json for Claude Desktop, or .claude/settings.json for Claude Code):

{
  "mcpServers": {
    "bigshield": {
      "command": "npx",
      "args": ["-y", "@bigshield/mcp"],
      "env": {
        "BIGSHIELD_API_KEY": "ev_live_your_key_here"
      }
    }
  }
}

Local development

If running from source:

{
  "mcpServers": {
    "bigshield": {
      "command": "node",
      "args": ["./packages/mcp/dist/index.js"],
      "env": {
        "BIGSHIELD_API_KEY": "ev_live_your_key_here"
      }
    }
  }
}

Environment variables

| Variable | Required | Default | Description | |----------|----------|---------|-------------| | BIGSHIELD_API_KEY | Yes* | — | Your BigShield API key. *Not needed for check_domain (public endpoint). | | BIGSHIELD_BASE_URL | No | https://www.bigshield.app | Override the API base URL (useful for local dev). |

Example agent interactions

"Is this email legit?" → Agent calls validate_email with the address, gets back risk score + recommendation

"Check all these signup emails" → Agent calls validate_emails_batch, gets a summary table

"How much of my quota have I used?" → Agent calls get_usage, reports remaining requests

"That email was actually fine, mark it" → Agent calls submit_feedback with false_positive