Building in the open
Why I am building PafCore
People pay for speed work. They still guess if it worked. That gap is why I started this.
You can try it yourself right now at pafcore.site. Paste any website address. You get a free report. No signup.
This piece is the simple version of why that page exists.
In simple words
When a website is slow, people leave. Shops sell less. Google cares about speed too. So businesses pay developers and agencies to make sites faster.
The usual free tools, like Google PageSpeed, give a long list of tips. The list is confusing. After you fix something, you still do not know if it truly helped, because the score jumps around on its own.
PafCore is my answer. It tells you what is slow, which file is the problem when we can find it, and what to fix first. Later, people can pay for stronger help: test a fix for real, get a ready file, or watch a site over time. The free report stays free.
What people run into
Someone opens PageSpeed or Lighthouse. They get a long list. Most items do not raise the score. They change a few things. They run it again. The number moves.
They still cannot tell if the fix worked, or if the page just wobbles from one run to the next.
Agencies and freelancers bill for this every week. Shop owners and product teams ask “did we get faster?” Developers and AI coding tools can write a change in minutes. Almost nobody can prove the issue is gone.
Why people already spend money here
This is not a guess. Google and Deloitte looked at about 30 million visits across 37 real brand websites. When a site got only a tiny bit faster, one tenth of a second, this is what they saw:
About 8% more sales on shop websites. About 9% more money per order. About 10% more bookings on travel sites.
Source: Google and Deloitte, “Milliseconds Make Millions.” That is why agencies keep charging for speed work.
AI tools can help write a fix quickly. They still cannot honestly prove the site got faster, because one lucky recheck is not enough. That gap is where PafCore earns its keep.
Who is already in the space
Free checks: Google PageSpeed Insights and Lighthouse. They are everywhere, and they are free.
Paid watch tools: products like DebugBear, Calibre, and SpeedCurve. They watch scores over time.
Big observability: Datadog, New Relic, Dynatrace. Heavy, and aimed at large teams.
AI helpers: Cursor, Claude, Copilot. Good at writing a fix. Weak at proving speed.
I am not trying to be “another list.” I am trying to be the clear report, and the proof people can trust.
Where we are today
- The free product is live. Anyone can run a scan on pafcore.site.
- Sign in and credits are started. You can buy a pack and spend credits on a lab test or an optimised image download when the finding supports it.
- Next: make the paid path so clear that strangers buy without me standing next to them.
Find the issue. Say what to try first. Prove whether that fix helped, without guessing from one noisy re run. Hand you something you can use, like a smaller image file, when we already made it.
If this sounds familiar
Run a free check on your own site. Keep the share link. Tell me if the top finding was wrong. That feedback matters more than a polite yes.
Try it on a real site
Paste a website address. Free report. No signup. Updates when something new ships are on the homepage.
Check my site