Every score on this site is produced by a published rubric, not by opinion, and never by commercial terms. This page explains exactly what the number means and, just as importantly, what it does not mean.
What the score measures
The score measures observable, checkable evidence. It asks how much a careful reader can verify about a tool from primary sources: the vendor’s own pricing page, its documentation, its changelog, and for open-source projects the GitHub API. It does not measure whether a product is enjoyable, well designed, or right for you. An automated pass cannot know that, and pretending otherwise would be dishonest.
How a score is built
A score has two parts. The first is evidence, worth up to 8.5: public pricing, documentation, maintenance signals, and how much we could actually verify. That part is what an automated pass can honestly establish, and it is capped, because a crawl cannot tell you whether a product is any good.
The remaining 1.5 comes from the people who use the tool: your star ratings, upvotes and follows. User ratings carry most of that weight, and a rating only counts at full strength once enough reviews exist that the average means something. One five-star review cannot mint a 10.
So a 10.0 requires both: a flawless evidence profile and a community that rates it highly. Every tool page shows the split, so you can see which half the number came from.
Hosted tools: how the 8.5 is split
- Pricing clarity, up to 1.5. Public pricing page, and whether we could parse real tiers from it. A page that exists but tells a reader nothing scores lower than one we can quote.
- Evidence depth, up to 2.0. Features, pros and cons, and a first-party screenshot. This is what separates a page we have genuinely examined from one where only the URLs exist.
- Free access, up to 1.25. A real free tier scores highest; a time-limited trial scores less; neither scores zero.
- Documentation, up to 1.25. Real developer documentation or API reference.
- Maintenance, up to 1.0. A public changelog is the strongest signal a product is still being worked on.
- Clarity, up to 0.75. Whether the vendor describes plainly what the thing does.
- API, up to 0.75. Whether a public API exists.
Open-source projects: how the 8.5 is split
- Adoption, up to 2.0. Stars, on a logarithmic scale so a very large project does not simply erase a smaller good one.
- Maintenance, up to 1.75. How recently the project was pushed to.
- Contributors, up to 1.5. Breadth of the contributor base, because a single-maintainer project is a real risk to depend on.
- Releases, up to 1.25. Whether tagged releases exist and are current.
- Licence, up to 1.0. A permissive OSI licence scores highest, copyleft slightly lower, an undeclared licence lowest.
- Issue hygiene, up to 1.0. Open issues relative to project size.
Figures come from the GitHub API and are refreshed daily.
What never affects a score
- Affiliate relationships. Whether we earn a commission is excluded from scoring entirely. It is not excluded from the order tools are listed in: see “How listings are ordered” below.
- Paid placement. We do not sell ranking. Any sponsored listing is labelled as sponsored, separately from affiliate disclosure.
- Third-party scores. We never merge another site’s rating into ours. Where we cite one, it is attributed and linked.
How listings are ordered
Ordering and scoring are two different things, and we keep them apart on purpose.
Every listing on this site, on category pages, the rankings table, the homepage, the alternatives block on a tool page and inside our articles, puts the tools we earn a commission on first. They carry an Our pick label wherever they appear, and the top three positions are marked with a medal. Everything below them is ordered by our score, descending.
What that placement does not do is change a number. The score printed next to an Our pick is the same score the rubric above produces from evidence, and it is displayed unmodified beside every promoted placement, so you can always see how far a tool has been moved and judge it yourself. Our top pick in a category is frequently not the highest-scoring tool in it, and we print both figures rather than hiding the gap.
Sort controls override this. Choosing “Highest score” or “Lowest score” on the rankings or browse pages gives you a pure score order with no placement applied.
Prices
A price is shown only where it has been verified by hand. Automated price extraction proved unreliable enough to be misleading, so unverified prices are suppressed across the site rather than published with a caveat. Where no price is shown, follow the link to the vendor.
Corrections
If something here is wrong, we want to fix it. Scores are recomputed when the underlying evidence changes, and a correction to a factual error will be made promptly.
Production services
We also sell video production work through our services page. It is delivered with a mix of our own pipeline and commercial software. Some of that software is reviewed on this site, some of it pays us a commission, and OpusClip is both.
Using a tool to deliver client work earns it nothing here. It does not change a score, and it does not change a position in any ranking, category page or comparison. The only way a commercial relationship affects placement is the one described above under “How listings are ordered”, and it is labelled wherever it applies.
This is checked rather than promised: our daily audit fails loudly if a tool in our production stack climbs the rankings without its score or its partner status having changed.