Most directories show you a score and ask you to trust it. We show you the ten numbers behind ours, on every tool page, including the one we calculate from reader behaviour rather than opinion.
The Short Version
Each tool is scored 1–5 on ten dimensions. The overall figure is the mean of all ten, rounded to one decimal place. Nothing is weighted and nothing is hidden — the full breakdown appears on every tool page, so you can see which dimensions carried the score and which dragged it down.
Nine dimensions are editorial judgements. The tenth is calculated from data.
The Ten Dimensions
- Capability: How much the tool can actually do, and how well it does the thing it exists for
- Ease of use: How quickly a competent person gets a useful result without reading the documentation
- Value: What you get relative to what you pay — including whether the free tier is genuinely usable or just a demo
- Reliability: Consistency of output, uptime, and how gracefully it fails
- Ecosystem: Integrations, APIs, extensions, and how well it sits alongside the other tools you already use
- Innovation: Whether it is still advancing, or coasting on an early lead
- Support: Documentation quality, responsiveness, and community depth
- Scalability: Whether it still works when the workload or the team grows
- Trust: Data handling, transparency, licensing, and company track record
- Demand (live): Calculated, not judged — see below
On every tool page these appear as a labelled bar chart, with a marker showing the average for that tool’s category. A 5/5 means little on its own; a 5/5 against a category average of 4.1 means something.
Demand Is Measured, Not Decided
The tenth dimension is not an opinion. Demand is computed from how often each tool’s page is actually read over a rolling 30-day window. Tools are ranked by readership, converted to a percentile, and mapped to a 1–5 score:
- Top 20% of tools by readership → 5
- Next 20% → 4
- Middle band → 3
- Lower band → 2
- Bottom band → 1
Where there is not yet enough traffic to rank meaningfully, every tool receives a neutral 3 rather than a fabricated figure.
Our own visits are excluded from the count, and the underlying view data is discarded after 35 days. Demand reflects what readers choose to look at — nothing else. It cannot be bought, and no commercial relationship influences it.
How Often Scores Change
Scores are recalculated every week, automatically. Every tool page displays the date its score was last reviewed.
The nine editorial dimensions change when a tool materially changes — a major release, a pricing shift, a licensing change, an outage or a security incident. Demand moves on its own as readership shifts, which means a tool’s score can change without us touching it.
Rated Is Not the Same as Hands-On Tested
These are two different things and we show them separately.
- The star rating reflects the ten dimensions above, assessed from documentation, pricing, release history, licensing and reader demand
- The “Hands-on tested” badge appears only where we have installed and used the tool ourselves. Those pages state the date, the version, and the hardware or plan we used
A tool can hold a high rating without carrying that badge. Most of the directory does not carry it yet, and we would rather say so than imply a depth of testing we have not done.
What a Score Does Not Mean
- It is not a ranking against every alternative. A 4.2 in a crowded category can be a better fit for you than a 4.7 in another
- It is not a recommendation for your specific situation. The “Best for” and “Limits” sections matter more than the number
- It is not paid placement. No tool can buy a score, a position, or inclusion in the directory
- It is not permanent. Scores move as tools and readership change
Independence and Affiliate Links
Some links to tools are affiliate links, meaning we may earn a commission if you sign up. This never affects a score. The nine editorial dimensions are set before any commercial relationship is considered, and Demand is arithmetic on reader behaviour that we cannot adjust.
Several tools we rate highly — including open-source ones such as Ollama, n8n and ComfyUI — have no affiliate programme at all and earn us nothing. That is deliberate. If the best answer pays nothing, it is still the best answer.
See our affiliate disclosure for the full position.
Corrections
If you believe a score is wrong, tell us what we have got wrong and why. We publish corrections rather than quietly editing pages.
Read our corrections policy or get in touch.
This methodology applies to all 195+ tools in the directory and is itself reviewed periodically. If something here is unclear, that is a fault worth fixing — please tell us.