TCP Ping

TCP-probe ping (not ICMP) — open a TCP connection to a host and port, measure how long the handshake takes, report whether anything is listening. Default port 443, optional host:port syntax. Bulk mode handles up to 100 targets at once.

Free to use, no signup. On the next page you confirm before we run the check.

What this check returns

Target: The IP address or hostname you submitted.
Resolved: IP address the hostname resolved to.
Port: TCP port that was probed.
Reachable: Whether a TCP connection to the port succeeded.
Latency: Time to complete the TCP handshake — a single sample in milliseconds, not an averaged RTT.

Bulk TCP ping (reachability check)

Find out whether a host accepts TCP connections on a given port and how long that connection takes to open. Submit up to 100 targets in one bulk check; the comparison table shows resolved IP, port, reachability and connect-time latency for each one.

Is this ICMP ping?

No. This is a TCP-probe ping — we open a TCP connection to the host:port and measure how long the three-way handshake takes. Classic ICMP ping needs raw sockets, which aren't available on shared hosting — and besides, lots of networks drop ICMP entirely while still letting TCP through, so this probe is a better answer to the question "is this service actually reachable?".

What this tool actually does

Parse target. Optional host:port — default port is 443 when none is given.
Resolve. If the target is a hostname, look up its first IPv4 A record. Raw IPs are used as-is.
Connect. Open a TCP socket to ip:port with a 4-second timeout.
Measure. Elapsed time from connect start to socket-open, in milliseconds. DNS time is not included.
Close. The socket is immediately closed; no protocol speak (no HELO, no GET, no TLS handshake).

Input accepted

example.com — default port 443.
example.com:25 — explicit port (SMTP, anything you want to probe).
8.8.8.8 — raw IPv4, default port 443.
8.8.8.8:53 — IP with explicit port.

What this tool returns

Target: The target you submitted, verbatim. Whatever you typed, including any :port suffix.
Resolved: The IPv4 address actually probed. For a hostname this is the first A record returned by the resolver — DNS round-robin pools only get one of the IPs checked per row.
Port: TCP port the probe targeted. 443 when you didn't specify one.
Reachable: yes when the TCP handshake completed, no when it failed, timed out, or was refused.
Latency: Whole milliseconds from connect-start to handshake- complete. A single sample, not an average — re-run if you need stability under load.
Error: Connect error message when reachability failed (timeout, refused, no route). Visible on the single-host detail card.

Common ports worth probing

22 — SSH
25 — SMTP
53 — DNS (TCP)
80 — HTTP
110 — POP3
143 — IMAP
443 — HTTPS (default)
465 / 587 — submission
993 — IMAPS
995 — POP3S
3306 — MySQL / MariaDB
5432 — PostgreSQL
6379 — Redis
27017 — MongoDB
8080 / 8443 — alt HTTP/S

Why use it?

Through-firewall reachability

A host that pongs back ICMP can still have its application port firewalled. TCP probe is the truthful answer to "can my client actually connect?".

Spot-checking new infrastructure

After spinning up a new server, run a quick set of host:port probes to confirm services are listening on the ports you expect.

Migration verification

After flipping DNS or moving a load balancer, confirm the new endpoint is up and roughly how long the handshake takes from our datacentre.

Latency comparisons

Bulk-check the same port across many hosts — useful for comparing region-to-region latency or detecting a slow outlier in a pool.

Quick uptime sanity check

Not a monitor — but for a one-off "is it up?" question across 30 hosts, this is much faster than 30 SSH sessions.

What this tool does NOT do

No ICMP. TCP connect only.
IPv4 only. Hostnames are resolved against A records; AAAA records aren't tried.
One sample per row. Not an averaged RTT — re-run for stability.
No protocol speak. We close the socket immediately after handshake; no HTTP GET, no TLS handshake, no SMTP EHLO. The port is reachable, but the service behind it might not actually answer.
No UDP. No DNS, no NTP, no SIP, no game-server probes.
No traceroute. Pure connect/disconnect.
4-second cap. Anything slower is reported as unreachable.

Bulk mode

Drop up to 100 targets in the textarea, one per line — mixing bare hostnames and host:port entries is fine. Each row is probed sequentially and the result table is exportable to CSV / JSON / YAML / XML. Good for post-deploy smoke tests, region-to-region latency surveys and pre-migration reachability audits.

Common failure modes

Could not resolve host — DNS failure on the hostname.
Connection timed out — no response within 4 seconds. Usually a firewall dropping packets silently.
Connection refused — host responded with TCP RST. Nothing is listening on that port (or the firewall is configured to actively reject).
No route to host — routing failure; host or network is unreachable.

Related tools

Pair this with SSL / TLS checker (cert details on the same host:port), HTTP headers (status + headers for HTTP services), the DNS records lookup (confirm what's resolving where), and domain & IP lookup (GeoIP / ISP / ASN on the IP behind the host).

Get started

Paste one or more targets in the form above — with or without an explicit port. Single submissions render as a detail card; bulk submissions as a side-by-side comparison table you can export.