Free tool · no sign-up

TTFB checker

Measure your site’s Time to First Byte from 15 locations on 6 continents at once. For each one you see where the time goes: DNS lookup, connection, TLS handshake and your server’s own response time.

The basics

What is TTFB?

Time to First Byte is the wait between asking for a page and the first byte of the answer. Nothing can render before it, so a slow TTFB makes everything after it slow too.

DNSConnectTLSServer
request startsfirst byte arrives = TTFB
Benchmarks

What is a good TTFB?

We rate the median of all locations, so one far-away location doesn’t decide it. Our probes sit in data centres on fast networks, so we rate the server’s answer on a stricter scale than Google’s, and show both.

≤ 200 ms

Excellent

A static or cached page, or one served from a CDN.

200 – 500 ms

Good

Fine for a dynamic page that asks a database.

500 – 1000 ms

Needs improvement

Worth a look: slow queries, no caching, a cold start or a server far away.

> 1000 ms

Poor

Visitors feel it before anything else has loaded. Fix this first.

Google’s web.dev threshold: good up to 800 ms, needs improvement up to 1800 ms, poor above. It is measured in real visitors’ browsers, slow mobile networks and redirects included, so it is looser. Every result shows it too.

How we measure

Fifteen fresh visits, timed phase by phase.

Each probe requests your page at the same moment as a new visitor would: new DNS lookup, new connection, nothing in the browser cache. When the address redirects, for example to https://www., we follow it, time the final page and tell you how long the redirects took. The Cache column shows whether your CDN answered from its cache (HIT) or passed the visit to your server (MISS, DYNAMIC).

DNS

Finding you

Looking up your domain’s address. Slow here means a slow DNS provider.

Connect

Reaching you

Opening the TCP connection: about one round trip to your server, so it grows with distance.

TLS

Securing it

The HTTPS handshake, one or two more round trips.

Server

Your answer

From sending the request to the first byte back: your application and hosting.

TTFB · Total

The sum

TTFB is all of the above. Total adds downloading the page and any redirects.

Fixes

How to improve TTFB

Read the columns: they tell you which fix applies.

“Server” high everywhere

Your back end is slow: cache pages, fix slow database queries, or move to faster hosting.

Fast nearby, slow far away

That’s distance. A CDN, or a server closer to your visitors, cuts Connect and TLS.

Behind a CDN but still slow

If Connect and TLS are tiny but Server is high, the CDN isn’t caching the page and every visitor waits for your origin.

DNS high

Use a fast DNS provider and keep your records’ TTL reasonable.

TLS high

Enable TLS 1.3 and session resumption, and serve a short certificate chain.

Redirect chains

Link straight to the final address, so visitors skip the extra round trips.

Google says 800 ms is good. Why do you rate stricter?

Google’s 800 ms is for what real visitors see in their browsers, on any network, phones on 4G included. Our probes measure from data centres, where the network adds little, so most of what you see is your server. For that, under 200 ms is excellent and under 500 ms is good. Each result gives both ratings.

Why is my TTFB different from Google PageSpeed?

PageSpeed measures from one place, often with cached connections, and lab tools add their own delays. We measure fresh visits from 15 places at once, so you see how far-away visitors experience your site.

Does the result change between tests?

A little. Networks and servers vary from second to second. For a trend, monitor the site: we record response time every minute from the same locations.

Can I share a result?

Yes. Every test gets its own link, kept for 7 days.

Is there a limit?

A few tests per hour from your network, and one fresh test per site per minute for everyone. We only send normal web requests to ports 80 and 443.