How to Add Sex Offender Safety Checks to Your AI Chatbot

AI chatbots are becoming the primary interface for safety-conscious applications — from neighborhood research tools to tenant screening platforms. This guide shows you how to connect your chatbot to live sex offender data using the Offenders.io MCP server.

The Problem

Users want to ask natural language questions like:

“Are there sex offenders near 123 Main St?”
“How safe is the area around this school?”
“Show me high-risk offenders within 2 miles of my home”

Without live data, your chatbot can only give generic advice. With Offenders.io, it can answer with real numbers from the national registry.

Option 1: MCP Server (Easiest)

If your chatbot is built on Claude, Cursor, or any MCP-compatible framework, just point it to our server:

https://mcp.offenders.io/mcp

Your chatbot instantly gets access to 6 tools:

Free: search_offenders_by_location, get_offender_stats, api_info — anonymized data
With API key: set_api_key, search_offenders_full, search_offenders_by_name — full records with names, photos, addresses

To unlock full records via MCP, just tell your chatbot “Set my Offenders.io API key to abc123”. No code changes needed.

Option 2: REST API + Function Calling

For chatbots built on OpenAI, LangChain, or custom frameworks that don't support MCP, use our REST API as a function/tool:

// Define as an OpenAI function
{
  "name": "search_sex_offenders",
  "description": "Search for sex offenders near a location",
  "parameters": {
    "lat": { "type": "number" },
    "lng": { "type": "number" },
    "radius": { "type": "number" }
  }
}

// When the function is called:
const res = await fetch(
  `https://api.offenders.io/sexoffender?lat=${lat}&lng=${lng}&radius=${radius}&key=${KEY}`
)

Example Conversation

User: Are there any sex offenders near Elm Street Elementary?

Chatbot: I found 12 registered sex offenders within 1 mile of Elm Street Elementary. 2 are classified as high-risk, 7 as moderate, and 3 as low-risk. The nearest is 0.3 miles away. Would you like me to expand the search radius or get more details?

Best Practices

Geocode first — convert addresses to lat/lng before calling the API. Use Google Maps, Nominatim, or any geocoding service.
Default to 1-mile radius — this covers the immediate neighborhood. Let users expand if needed.
Summarize, don't dump — give counts and risk breakdowns first, then offer details on request.
Cache results — offender data updates daily, so caching for a few hours is safe and saves API calls.
Use fuzzy search — when searching by name, enable fuzzy=true to catch misspellings and partial matches.

Pricing for Chatbots

Most chatbot use cases fall under 2,000 calls/month ($0.20/call = $400/month max). The free tier covers 50 calls/month for development. For high-volume chatbots, enterprise pricing starts at $0.15/call. See pricing →

Connect Your Chatbot

Start with the free MCP server or get an API key for full access.

MCP Server Docs Get API Key

How to Add Safety Checks to Your AI Chatbot