Private landing page teardown
Your page is not ready to be public.
So why is asking for feedback?
Paste a URL. Foldscore opens the page in a real browser, reads it the way a cold visitor does, scores it out of 100, and rewrites the copy that is losing you the click. About a minute, no signup, and nothing is posted anywhere unless you tick a box.
No signup · 3 free a day · private unless you say otherwise
Median across 8 pages: 70 out of 100 · what each band means
Why this exists
Four reasons, and none of them is “AI”.
It is private, and that is not a small thing
Every free way to get your landing page reviewed requires posting the URL somewhere public, permanent and indexable. Reddit, Indie Hackers, X. If you are in stealth, mid repositioning, or testing a price you have not committed to, that is not a small tradeoff. It is the whole reason you did not ask.
Here the report lives on an unlisted link that search engines are told to skip. The gallery is a checkbox, off by default. Nobody sees your page unless you send them the link.
You can run it ten times
A human teardown is a one-shot purchase. You buy it, you read it, and then you feel faintly rude asking the same person to look again after you changed three lines. Nobody gets their page roasted ten times in a week.
10 credits is the point, not the limit. Run it after every rewrite and watch the score move. Run it on the two competitors you keep losing to. Every re-run is diffed against your last one, dimension by dimension, so you can see whether the change did anything.
It rewrites. It does not just grade.
Handing back a number is the easy half. Every scorer in this category stops there, and a score with no replacement copy is a to-do item, not a fix.
You get three rewritten hero blocks on three different angles, replacement button labels, and the objection map that says where each silent doubt has to be answered. Written from your page’s own words. Where a line needs proof your page does not carry, it comes back as a marked slot to fill, never an invented statistic.
You cannot A/B test your way out of this
Detecting a 20% improvement on a 3% baseline conversion rate needs 27,827 visitors for one two-arm test at conventional significance and power. At 500 visits a month that is 4.6 years to learn one thing.
That is arithmetic, not opinion, and it is why expert heuristic review is the standard prescription for low-traffic pages. CXL says the same thing. You are not short of tests. You are short of traffic to run them with.
What gets scored
Seven questions a visitor answers without noticing.
Five come from reading the page. Two come from looking at it, because a browser opens your page and takes a screenshot before anything is scored. When the screenshot fails, the report drops to five and says so rather than guessing.
5-Second Clarity
Does a stranger know what this is and who it is for within five seconds?
Value Proposition
Is the promise specific enough that a competitor could not paste it on their own site?
Call to Action
Is there one obvious next step, and does it feel low-risk to take?
Trust & Proof
Is there evidence that other people use this and it works?
Friction & Objections
What unanswered question stops a warm visitor from converting?
Visual Hierarchy
needs eyesDoes the eye land on the headline and then the primary action, and is that action visually dominant?
Above-the-Fold Completeness
needs eyesCan a visitor understand what this is and act on it without scrolling?
One number
They roll up into a score out of 100 and a letter grade, so you can tell in one glance whether this page is worth more traffic or a rewrite.
Straight about the limits
What this is not.
Not behaviour analytics
No heatmaps, no session recordings, no funnel charts. Those tools tell you what people did. This tells you what a stranger understands. Different question, and you need traffic before the first one has an answer.
No install, no traffic required
Nothing to add to your site. No script, no tag, no account on your end. It works on any public URL, including one that has never had a visitor, and including pages that are not yours.
Not better than a good human
For one deep look at one page, a $99 to $350 video from someone who has done hundreds of them is probably better than this. That is an honest read of the market. What they cannot do is come back in four minutes when you have changed the headline.
No conversion-lift promises
You will not find a percentage on this site claiming what Foldscore does to your conversion rate, because measuring that would take the same 27,827 visitors as everything else. What you get is a specific, checkable critique and the replacement words.
Opted in
Reports people chose to publish.
Reading someone else’s is the fastest way to calibrate what a 40 looks like next to an 80. Yours only appears here if you tick the box.
hey.com
2026-07-27
The page nails the what: 'We finally fixed your email + calendar.' It fails the why. A visitor learns HEY exists and that famous people like it, but not what problem it solves for them or why they should care more than they care about Gmail. The social proof is strong but names a category of people, not a category of results.
basecamp.com
2026-07-27
The H1 'The refreshingly straightforward project management system that's rock-solid and easy to use' is a category descriptor, not a value promise. It tells the visitor what Basecamp is, not what Basecamp does for them. The page then spends 651 words on testimonials and company history instead of answering the cold visitor's real question: why is this better than the tool I already use?
web-production-fd8c4.up.railway.a…
2026-07-27
The page nails what it does, "returns a score out of 100 across five dimensions", but buries why a founder should care. The Notion teardown example proves the tool works, but the page offers no quantified proof that FirstFold users actually improve conversions or that the scoring framework is reliable. A cold visitor lands knowing this audits pages, not whether it's worth their time.
What costs money and what doesn’t
The free score is the real score.
You get the number and the diagnosis without paying or signing up. Credits buy the part that would otherwise cost you an afternoon of writing.
Free
No account- The full score out of 100 and the letter grade
- Every dimension score, each with a one-line verdict quoting your page
- The single highest-priority fix, in full
- An unlisted link with a score card, private until you share it
- 3 pages per day
Foldscore Pro
$39 for 10- 10 credits that never expire. One credit unlocks one full report, and re-reading it is free forever.
- 3 rewritten hero variants: headline, subhead and CTA, on three different angles, written from your own copy
- The complete issue list, ordered by impact per hour of work
- CTA rewrites, and an objection map saying where each doubt gets answered
- A one-click prompt for Cursor or Claude Code on every fix
- Not useful? Ask for a refund and you get it.