Is a port reachable from the internet? Connects like telnet and shows what answers.
A port is open when something is listening on it and nothing between you and the server stops the connection. The question comes up after every firewall change: did the rule open what it should, and is anything open that should not be? Checking from your own network cannot answer it, because your office sees the server differently from the internet.
Enter a public host name or address and a port. The tool opens one TCP connection from our server in India, the way a telnet client does, reports whether it was accepted, refused or ignored and how long it took, then listens for the greeting the service sends. A web server that stays silent is sent a single HEAD request. Nothing is ever sent to guess a password, and only one port on one host is tested per request.
Opening a connection to someone else's server is the same act an attacker starts with. Only test hosts you own or are permitted to test. The check makes a single connection, never guesses passwords, and is rate limited, but the confirmation is a reminder of whose server it is.
Your office may be on an allow list. Many firewalls open a port only to named addresses. That is good practice, and this tool sees it as the public does: closed or filtered.
A firewall that drops traffic gives no answer, so we wait five seconds before calling it filtered. A closed port answers instantly with a refusal.
No. Private, loopback and cloud-metadata addresses are refused, and so are names that point to them. Our server is on the internet, so it can only test what the internet can reach.
Not necessarily. Cloud providers commonly block outbound and inbound port 25 by default, and many mail servers accept connections only from specific networks.