WP·VILLAIN
Performance Documented

Measuring WordPress TTFB From the Command Line: curl, wget, and Why to Test as Googlebot Too

October 28, 2026 · wordpress ttfb performance curl wget caching bash

Time to First Byte is the number I check first when a WordPress site feels slow. It needs no browser, no Lighthouse run and no account, only curl and a URL. It also tells you whether the problem sits on the server or in the front end, which decides where you spend the next hour.

What counts as good

Google’s web.dev guidance on TTFB puts it like this: TTFB comes before First Contentful Paint and Largest Contentful Paint, so a slow server response caps every metric after it. The rough thresholds:

TTFB Rating
0.8 s or less Good
0.8 to 1.8 s Needs improvement
Over 1.8 s Poor

These are targets for the 75th percentile of real visitors. One curl run from your laptop is a sample, not a verdict.

One curl command

curl -w "\nTTFB: %{time_starttransfer}s\n" -o /dev/null -s -D - https://example.com

-o /dev/null throws the body away, -s hides the progress meter, -D - prints the response headers to the terminal, and -w appends the timing. You get the headers and the number in one go:

HTTP/2 200
content-type: text/html; charset=UTF-8
x-litespeed-cache: hit
server: LiteSpeed

TTFB: 0.762227s

The headers matter as much as the number. A x-litespeed-cache: hit or an x-cache header tells you whether you measured the page cache or PHP. Measuring a cache hit and calling the site fast is how slow uncached pages stay hidden.

To see where the time goes, ask curl for each phase:

curl -o /dev/null -s -w "dns %{time_namelookup}s  connect %{time_connect}s  tls %{time_appconnect}s  ttfb %{time_starttransfer}s  total %{time_total}s\n" https://example.com

If dns, connect and tls are small and ttfb is large, the server is spending the time building the page. If the first three are large, you are looking at the network path or the TLS setup, not WordPress.

wget as a second opinion

wget has no timing variable, but it is often installed where curl is not:

wget --server-response --spider --timeout=30 https://example.com 2>&1 | grep -E "(HTTP|Length|time)"

For a rough total, wrap it in time:

time wget --spider --quiet https://example.com

That is total request time rather than TTFB, so use it as a sanity check only.

Run it more than once

The first request after a cache purge is slow, the next ones are fast, and any single number is noisy. I run five requests and average them. I wrote that into a script in wp-ops, which also prints DNS, connect and SSL timings and rates the result against the thresholds above:

wp-ops ttfb-test https://example.com

It runs from your machine, so your own distance to the server is part of the result. For the server’s own view, run the request on the server over SSH. The second script does that, and it does it once per URL for four user agents: the default, Googlebot, AhrefsBot and Screaming Frog:

wp-ops remote-ttfb-ua web@example.com https://example.com/ https://example.com/blog/

Why test as a bot

A WAF, a rate limiter or a cache rule can treat crawlers differently from visitors. A cache that keys on a cookie bots never send, or a bot-protection layer that challenges unknown user agents, will serve visitors cached pages while crawlers get the uncached path. Every manual check in your browser looks fast, and the slowness only shows up in Search Console’s crawl stats.

If the default user agent is fast and the bot user agents are not, look at cache bypass rules and bot-protection settings before you touch PHP.

What to do with the number

  • Cache hit, fast: the server side is fine. Look at images, scripts and fonts next.
  • Cache hit, slow: a network or TLS problem, or an overloaded server. Check the connect and tls phases.
  • Cache miss, slow: PHP and the database are the cost. Object caching, a lighter plugin set and PHP-FPM worker settings are the levers.
  • Fast for visitors, slow for bots: a cache or WAF rule, as above.

The broader checklist of what to change once you know the server is the problem is in WordPress Speed Optimization Guide 2025.

Measured something odd with these commands? Find me on Mastodon at @jfrumau@mastodon.social.