Use seecrime in Claude and ChatGPT

seecrime works inside Claude and ChatGPT. Once you connect it, you can ask the assistant about reported crime in plain English, and it answers by looking things up in seecrime: the same incidents, counts and risk scores as the map, for the 41 US and Canadian cities we cover. The connection can only read. It cannot change anything in seecrime.

What you can ask

The assistant has 7 read-only seecrime tools to work with. These are the kinds of questions they answer:

For a named place, such as a station or a market, the assistant works out where it is itself and asks seecrime about the area around that point. seecrime does not look up landmarks: it finds only US street addresses, and only for a risk score.

It works with what the cities publish. Some feeds run days or weeks behind, some publish dates without times, and where no data is held for a period the assistant is told to say so rather than report zero. A risk score by street address works for US addresses only; in the Canadian cities the assistant can score a point from its map coordinates instead.

The connection does not draw maps, write the PDF or Word reports, or see your saved zones. Those stay in the seecrime app.

Who can use it

Any paid seecrime plan includes it: Pro, Team or Enterprise. You connect by signing in with your seecrime account, and the assistant then reads data as you. A free account cannot connect: if you try, seecrime.com shows you why, with a link to the plans.

Claude and ChatGPT need nothing installed: you add seecrime in their settings. Custom connectors depend on your Claude or ChatGPT plan and settings, so at work the person who manages your account may need to allow them. If your seecrime account is on a paid plan and the app still says it cannot connect, contact us.

How to connect

Every app uses the same address: https://www.seecrime.com/mcp. Add it once as a custom connector, sign in to seecrime.com when the app sends you there, and approve the app. If an app's screens look different from the steps below, it may have changed them since we checked.

Claude: claude.ai, Claude Desktop and the mobile apps

1. Go to Customize → Connectors and click Add custom connector. On a Claude Team or Enterprise plan, an Owner adds it once for the organization instead, under Organization settings → Connectors → Add → Custom.
2. Enter the URL https://www.seecrime.com/mcp and click Add. If the dialog asks how Claude identifies itself, keep Use Claude's published identity; you need no client ID or secret.
3. Click Connect on the connector, sign in to seecrime.com in the window that opens, and approve the app.

A connector added on claude.ai is there in Claude Desktop and the mobile apps too, signed in as the same account.

ChatGPT

1. Open ChatGPT Plugins, select the plus button, then Add custom MCP server. Your workspace's policies decide whether this is available.
2. Give it a name, enter the URL https://www.seecrime.com/mcp, and choose OAuth for authentication. ChatGPT identifies itself with its own published client details; you need no client ID or secret.
3. Create it, then sign in to seecrime.com when ChatGPT sends you there, and approve the app.

Claude Code

claude mcp add --transport http seecrime https://www.seecrime.com/mcp

Then run /mcp in Claude Code: the server shows as needing authentication, and choosing it opens your browser to sign in and approve. From a shell, claude mcp login seecrime does the same (add --no-browser on a machine without one); claude mcp logout seecrime signs out.

Other apps

Using another app, such as Cursor, VS Code or Gemini CLI, or connecting a script with an API key? The developer guide has the steps for each.

About the risk score

A risk score ranks one point against the other places in its city by reported crime over the last two years, and says how it compares with the city average, by type of crime, with an estimate of its uncertainty. Scores compare places within a city, not cities with each other. Where seecrime holds no data for a place, the assistant gets no score and is told that this means no coverage, not that the place is safe. Every score comes with this statement, and the assistant is told to show it with the score:

Privacy and what we record

• Read-only. The connection can only look up published crime and 311 data through these tools. It cannot see or change your account, your saved zones or anything else in seecrime.
• Your conversation stays with the AI app. seecrime receives only the lookups the assistant makes, such as a city and a date range, or an address to score, and sends back the results. What the app does with those results is covered by its own terms and privacy policy.
• What we log. Our MCP log has one line per request the app sends, naming the tool and the connection, never what was looked up: an address is personal data. Like every request to the site, each one also appears in our server's request log, with the path (/mcp), the status and how long it took; that log does not record your IP address. The lookup itself travels in the body of the request, which neither log records. To score an address, seecrime sends it to the US Census Bureau geocoder to find it on the map, and the point to the other services the risk score uses; all of them are listed in the privacy policy.
• Third-party data. Addresses, offense descriptions and 311 request text come word for word from city and police feeds. We do not alter them, and the assistant is told to treat them as data, not as instructions.
• Disconnect any time under Profile, Connected AI apps. Changing or resetting your password, leaving your organization or moving to the free plan disconnects every app at once.