About
Built because the data was there and the history wasn't.
FrictionScope is page friction intelligence for builders. It snapshots Microsoft Clarity's per-URL friction signals every day (rage clicks, dead clicks, quickback clicks and scroll death), classifies each page into a problem archetype, and serves the verdict as read-only Model Context Protocol tools inside your editor.
Why it exists
Microsoft Clarity already records the signals that tell you a page is failing: people clicking something that does not respond, clicking the same thing in frustration, landing and immediately hitting back, never scrolling past the hero. It gives that data away for free, and it exposes it programmatically through its Data Export API.
That API has one hard constraint: it only returns the previous one to three days. Ten requests per project per day, a thousand rows, no pagination. You cannot ask it what a page looked like last month, and you cannot ask it whether the change you shipped on Tuesday helped.
To be precise about it, because the distinction matters: Clarity does not forget. It keeps recordings for 30 days and heatmaps for 9 months, and its own dashboard will show you them. What is capped at three days is machine-readable access. So per-page history and before-and-after comparison are possible, but only if someone takes a snapshot every day and keeps it.
That is the entire product. One request a day, stored forever, turned into a verdict.
How it works
You connect a site by pasting a Clarity API token you generate yourself and can revoke at any time. Once a day we pull that project's metrics broken down by page URL and store the rows with the date.
Every stored page then gets classified. Not into a score, into a named problem: a page with real traffic and a high rage-click rate is Broken but busy; a page people land on and immediately leave is Disappointed; a page throwing JavaScript errors is Actually broken. A page with too little traffic to judge is labelled Insufficient data rather than guessed at.
Because every snapshot is dated, two things become possible that the API alone cannot do: you can see whether a page is getting worse, and you can compare the window before a change against the window after it.
What it deliberately is not
Not a heatmap or session-replay viewer. Those are UI-only in Clarity and are not in the export API at all. We surface the counts behind them, which is the part you can act on from an editor anyway.
Not a general analytics dashboard. There are enough of those. FrictionScope answers one question: which pages should you fix, and did your fix work.
Not real-time. Daily is the right cadence for "what should I fix next", and the API's rate limits would make real-time impossible regardless.
Not affiliated with Microsoft. Clarity is a trademark of Microsoft Corporation. FrictionScope reads your own Clarity data over a token you control.
Who makes it
FrictionScope is an independent product, built and run by a solo maker alongside a small family of other tools: QueryScope for Search Console data in your IDE, and AppScreenshotStudio for App Store screenshots, among others.
It runs on a single server in Germany. There is no venture funding behind it and no growth team. If something is broken or wrong, the person who wrote it is the person who reads the email.
What it costs
Nothing, currently. FrictionScope is early, and there will be no paid plan until the classifier has been calibrated against enough real snapshots to produce verdicts worth paying for.
If that changes, existing accounts get at least 30 days' notice by email, and nobody is charged without explicitly picking a plan first. That commitment is in the Terms, not just on this page.