Is this still climbing, or am I late?
Every research tool in this category gives you a number. None of them gives you the derivative. MurmPulse measures the change between readings you took yourself, prints the sample count beside every figure, and refuses to tell you what happens next.
Measured, not modelled. Every figure comes from readings you took, on your machine. Where a number cannot be known from the page, MurmPulse prints an em-dash and the reason instead of estimating it.
How it works
- Open a product page and click the icon. That records one reading of whatever the page publishes about itself — price, sold count, review total, rating.
- Come back and click again. Days later, weeks later; the schedule is however often you were going to look anyway.
- From the third reading it reports which way the figure moved, at what rate per day, and how many readings that rests on.
What it refuses to tell you
Every MurmTools extension refuses at least one number that other tools invent. This one refuses the future.
- No forecast. No projection, no “on track for”, no dashed continuation of the line. There is no field in the code for one.
- No trend below three readings. A line through any two numbers has a direction. Calling that a trend would be inventing information.
- No percentage from zero. There is no percentage change from a baseline of zero, and Infinity rendered in a tile is a lie with a confident typeface.
- No figure when the page reports a floor. A page that says “500+ sold” is declining to give a count. Three readings of the same floor is not evidence that nothing moved, so the change is refused rather than printed as zero.
- No smoothing across a gap. If you did not look for three weeks, the record shows three weeks of nothing, because that is what was measured.
The first visit
A velocity tool has no history on day one, and that is where this kind of product usually starts lying. So on the first reading MurmPulse takes whatever the page already says about its own past — a sold count beside a listing date gives an average per day since it was listed — and labels it exactly that. An average since launch is not this week’s rate, and the panel does not let the two be confused.
Why it needs no server
Your own repeat visits are the sampling schedule. A hosted tool would have to poll every product for every user to build the same history, which is expensive — which is why the category sells you a point estimate instead. A local tool gets its time series for free.
What it stores, and what it does not
The readings, on your machine, and nothing else. History is kept at full resolution for a month, daily for a year and weekly after that, so it stays inside a budget rather than growing forever — and the panel tells you when older readings have been dropped. There is no account, no server, no analytics, and no network code of any kind in the package.
Free, with no paid tier. Three permissions — activeTab,
scripting, storage — and no host permissions at all.
Questions
How can I tell if a product is still selling well or already saturated?
Take a reading each time you visit. From the third, MurmPulse reports which way the figure moved and at what rate per day, always with the number of readings and the window it rests on. It cannot tell you what happens next, and does not claim to.
Why three readings?
A line through any two numbers has a direction. Below three, MurmPulse shows the readings and says how many more it needs.
Does it predict future sales?
No, and there is no field in the code for one. A trend here is a record of what happened between readings you took.
Is it free?
Yes, with no paid tier, no account and no server.