—
—
We couldn't read the page.
This is a pilot. If the error keeps coming back, get in touch — I'll fix it personally.
—
—
—
—
—
—
These are still on your page, right now. Let me fix all of them — I look myself, prioritise, deliver.
I'll fix the findings above and dig up the rest — one cold-read, one senior designer, site solved.
book Gustaf rather talk first? drop me a line →Three specialist agents — just them. craftsman reads craft + a11y, translator picks apart copy + slop, scout checks market + facts. No Gustaf on the free scan — I don't have a notification system for every single run, so the agents did the work here. None of them keep your URL after the scan.
The agents don't always agree. What's shown here is what all three line up on — that's why it says confidence 0.86, not 1.0. What's left is counted directly on the page, not guessed.
Three concrete things, measured straight from markup: contrast violations (WCAG AA), extra font variants beyond the three you already use, and heading hierarchy skips (h1 → h3 with no h2). No guesses — if the count is zero, it's zero, and we hide the block.
We book 20 minutes, I ask five questions, then I run /cold-read: five agents in a pipeline plus my hands-on pass, my techstack and my delivery. You get a page that converts better, the problems fixed, evidence for every change. Fixed price, fixed delivery — no hour estimates.
Pilot. The scan is real, the agents are real, the price is real — but it's early. If anything looks off, get in touch and I'll look myself.
Honestly — did this feel brilliant? Rubbish? Want to see what didn't make the cut? I read every email myself. No newsletters, no sales reps — just me.