How we rate
One verdict. Four signals. No invented numbers.
Every review ends in one of three verdicts. Under each one we publish the four things we can actually check without installing the extension — and the exact thresholds that turn those facts into words.
The three verdicts
A verdict answers one question: would we install this for the job it is built to do? It is a judgment, made by a person, from the public record. It is not calculated.
| Verdict | What it means | Reviews |
|---|---|---|
| Recommended | The public record supports choosing it for the job it is built to do, and we would install it ourselves for that job. | 52 |
| Worth it, with caveats | Genuinely useful, but something specific belongs in your decision — thin ratings, a long gap since the last update, a hard paywall, or wide page access. Every review at this level names the caveat next to the verdict. | 21 |
| Not our pick | We found a better option for the same job, or the public record is too thin to stand behind. It is not a safety warning and not an accusation — it means look at the alternatives first. | 6 |
Those counts are live — the actual split across all 79 reviews on the site right now. We publish them because a scale where nearly everything lands in the top tier tells you nothing, and you should be able to check that for yourself rather than take our word for it.
Every verdict that is not “Recommended” must name its reason. One short sentence sits next to the badge saying the specific thing that held it back — a thin rating sample, a long gap since the last update, a paywall, a data disclosure. A caveat you cannot see is not a caveat; it is a mood.
The four signals
Every review shows the same four readings, taken from the Chrome Web Store listing cited in its sources. They describe the evidence available about an extension, and they are shown so that you can disagree with us: if you weigh a small user base differently than we do, the number is right there to reweigh.
Adoption — how many people run it
Read from the public user count on the listing. Listings that publish no count are marked “Not published” rather than guessed at.
| Public users | We call it |
|---|---|
| 1m or more | Very widely used |
| 100k – 1m | Widely used |
| 10k – 100k | Established |
| Under 10k | Small user base |
A small user base is not a mark against an extension’s quality. It means there is less public evidence about it — a different thing, and worth knowing before you trust it with your drafts or your passwords.
Rating — the store score, in context
A raw star rating means very little on its own, so we do two things to it. First, if a listing has fewer than 100 ratings we refuse to read anything into the score at all — a 5.0 from eight people is not evidence. Second, above that floor we compare the score to the median rating of every extension we review in the same category, because 4.2 means something different among password managers than among Pomodoro timers.
| Condition | We call it |
|---|---|
| Fewer than 100 ratings | Too few ratings to weigh |
| 0.2 or more above the category median | Above category average |
| Within 0.2 of the category median | About category average |
| 0.2 or more below the category median | Below category average |
Upkeep — how recently it was updated
An abandoned extension is a real risk: browser APIs move, and unmaintained code is where compatibility and security problems collect. We measure the gap between the extension’s last store update and the day we checked the listing — not today’s date. That way a page built months later never quietly claims an extension is fresher or staler than it was when we actually looked.
| Gap at our fact-check date | We call it |
|---|---|
| 0–6 months | Actively maintained |
| 6–18 months | Updated less often |
| More than 18 months | No recent updates |
Cost — what you actually pay
Taken from the developer’s own pricing page rather than the store’s “free” label, which is true of almost every extension and therefore tells you nothing.
| Model | We call it |
|---|---|
| free | Free |
| freemium | Free tier, paid upgrade |
| paid | Paid |
A worked example
This is the real block from our Web Developer review, drawn by the same component that renders it there — strong adoption, a solid rating, and a long gap since the last update, with a verdict that says so out loud.
Our verdict
Worth it, with caveats
Why: No store update since June 2024 — verify modern compatibility first.
- Adoption
- Very widely used1,000,000 users
- Rating
- About category average4.5 / 5 from 2.8K ratings
- Upkeep
- No recent updatesLast updated June 28, 2024 — 26 months before our check
- Cost
- Free
Version 3.0.1 · Store facts verified Aug. 20, 2026 · Not hands-on benchmarked
What this system does not do
- We do not hands-on benchmark. No timing runs, no accuracy scoring, no controlled side-by-side installs. If that ever changes, this page changes first.
- The signals do not add up to the verdict. There is no formula, deliberately. An extension can be strong on all four signals and still be the wrong tool for a job; another can be weak on adoption and still be the right pick because nothing else does what it does. The signals show the evidence — a person still makes the call, and names the reason whenever the call is not a straightforward yes.
- Signals are dated, not live. Each reflects the listing on that review’s fact-check date. Store metrics move; check the listing before you install.
- We do not audit code or network traffic. The permissions and privacy sections summarise what a listing and policy disclose. We do not decompile extensions or monitor what they send.
- No numeric score, ever. Not out of five, not out of a hundred. A number would imply a precision this method does not have.
Common questions
Why is there no score out of ten?
Because we do not hands-on benchmark extensions. A number like 8.4 implies a measurement we did not take. Everything here is either a fact copied from a public listing or a judgment clearly labelled as one.
Does “Not our pick” mean an extension is unsafe?
No. It means we found a better option for the same job, or that the public record is too thin to stand behind. It is not a security finding. If we believed an extension were dangerous, the review would say so directly.
Can a developer pay to change a verdict?
No. There is no paid placement, no sponsored review, and no affiliate relationship anywhere on this site.
How often do the signals change?
They are read from the store listing on the fact-check date printed on each review, and change only when we re-verify that listing. Nothing updates silently in the background.
Related
For where our facts come from and how sources are prioritised, see the review methodology. For corrections, conflicts of interest, and independence, see the editorial policy. To run the same checks yourself on an extension we have not covered, use the vetting guide.