AI Integration · MCP Server

Bring HostTracker monitoring into your AI assistant

Connect HostTracker over the Model Context Protocol and run instant website checks from 300+ global locations — right inside Claude and other AI assistants. Ask whether a site is up, resolve DNS, look for blacklist listings, trace routes and more, in plain language.

Uses your HostTracker API token · Works with Claude Code, Cursor & config-file MCP clients
hosttracker · run_instant_check
Youis example.com up right now, checked from Europe and Asia?
HTRunning an HTTP check on example.com from HostTracker's network…
London, UK200 · 11 ms
Frankfurt, DE200 · 9 ms
Singapore, SG200 · 14 ms
Tokyo, JP200 · 12 ms
HTUp across every region — all four locations return 200 OK.
What you can do

Real checks from your AI, in plain language

HostTracker's MCP server exposes your assistant to the same instant-check engine behind the homepage tool — now driven by conversation and run from 300+ worldwide locations.

Availability

Up anywhere in the world

Check HTTP/S status and response time for any URL from 300+ locations, so you see whether it's down everywhere or only in some regions.

Connectivity

Ping, port & traceroute

ICMP ping, TCP port reachability, and traceroute from multiple countries — find where connectivity actually breaks.

DNS & reputation

DNS & blacklists

Run a DNS query, or check whether a domain or IP is on a DNS blacklist (DNSBL) or the Russian registry blacklist.

Safety

WHOIS & Web Risk

Look up WHOIS registration details, and check a site against Google Web Risk for phishing and malware flags.

Output

Per-location results

Every check returns structured, per-location results your assistant can read and summarize — status, timing, and errors.

Compatibility

Works with your AI client

Connect from Claude Code, Cursor, the mcp-remote CLI, or any config-file MCP client, using your API token.

Saved monitors & uptime history live in the dashboard
Those stay in HostTracker for now — MCP focuses on ad-hoc checks, with account access on the roadmap.
Open dashboard →
Connect in minutes

Add HostTracker to your MCP client

No local server to run — point your MCP client at the hosted endpoint with your token.

{
  "mcpServers": {
    "hosttracker": {
      "type": "http",
      "url": "https://mcp.host-tracker.com/mcp",
      "headers": { "Authorization": "Bearer YOUR_TOKEN" }
    }
  }
}
1

Mint a token

On your Profile page, open the "API v2 Quota & JWT" card, check the ic scope, and create a token.

2

Add the endpoint

Copy the config above into your MCP client (e.g. .mcp.json) and replace YOUR_TOKEN with your token.

3

Restart the client

Reload your MCP client so it discovers the tools: run_instant_check, get_check_result, list_check_types.

4

Ask away

Use natural language to run checks and read the results — a one-click connector for Claude Desktop is planned.

Example prompts

Things you can ask

Just describe what you want to know — your assistant picks the right check and runs it.

Availability & speed
Check if example.com is up right now from Europe and Asia.
Check example.com from the US and show response time per location.
Is my site loading everywhere, or only in some regions?
Network & DNS
Run a traceroute to example.com from multiple regions.
Is port 443 open on example.com from different countries?
Is example.com's IP on any DNS blacklists?
Security & lookups
Check example.com for Google Web Risk / malware flags.
Do a WHOIS lookup on example.com.
What instant-check types can you run?
Model Context Protocol

What is MCP?

MCP is an open standard for connecting AI assistants to external tools in a uniform way — increasingly supported across AI assistants and agent frameworks. Instead of copy-pasting between a monitoring dashboard and your AI, MCP lets the assistant call HostTracker directly and reason over the results. Learn more at modelcontextprotocol.io.

Frequently asked questions.

Any MCP-compatible client that supports a custom Authorization header: Claude Code, Cursor, the mcp-remote CLI, and other config-file MCP clients. A one-click connector for Claude Desktop / claude.ai is planned (it requires OAuth) and is not available yet.

You can run one-off instant checks of any URL you name - HTTP/S availability, ping, TCP port, traceroute, DNS query, DNS blacklist (DNSBL), Russian registry blacklist, WHOIS, Google Web Risk, site crawl, and page-load waterfall - from HostTracker's 300+ global locations, and read the per-location results in your chat. It does not (yet) manage your saved monitors or query uptime history; those run in the HostTracker dashboard.

You pass a per-user HostTracker API token as an Authorization: Bearer header in your MCP client config. Mint one on your account Profile page, under the "API v2 Quota & JWT" card, with the "ic" scope. Your plan must include API access. All quota and rate limits are enforced by HostTracker against that token.

It is effectively read-only for your account: the tools start instant checks and return results. They cannot create, edit, or delete your saved monitors, contacts, or settings - there is no write access to your account through the MCP server.

Yes. Add an http MCP server entry pointing at https://mcp.host-tracker.com/mcp with your Authorization: Bearer token in the headers, and the tools appear as hosttracker_run_instant_check, hosttracker_get_check_result, and hosttracker_list_check_types.

MCP is an open standard that lets AI assistants call external tools in a uniform way, supported across Claude and other assistants and agent frameworks. HostTracker's MCP server implements it so your AI can run monitoring checks directly, instead of you opening a browser.

Free trial available now

Put your monitoring in your AI assistant

Create an account, mint an API token, and connect HostTracker to Claude in a couple of minutes.