How we score — Folium

Every Folium score — whether it lands on a technical assessment or a hands-on review — comes from the same five-dimension rubric, weighted and totalled the same way. We publish it here because a number without a method is just an opinion wearing a costume.

the five dimensions

30% Does the job Does the product reliably do what it claims? This is the core functional check — not benchmarks for their own sake, but whether it holds up against its own promises in real use.

20% Value Total cost — including subscriptions, licensing tiers, and any recurring fees — weighed against comparable alternatives. A good product at a bad price scores lower here, not on "does the job".

20% Freedom & exit Sylvaris's distinctive axis: lock-in, open standards, data export, self-host friendliness. How easy is it to leave, or to move your data elsewhere, if you change your mind? Lock-in is treated as a real defect, not a footnote.

15% Ease & experience Interface design and day-to-day usability, setup difficulty, documentation quality, and ongoing maintenance burden. If the UI fights you every day, that costs points here regardless of how capable the product is underneath.

15% Longevity Update track record, community or vendor health, and how long you can reasonably expect it to be supported. A product that might be abandoned next year scores lower even if it's excellent today.

score anchors

Each dimension is scored 0–10 before being weighted into the published total, using the same anchors throughout — so a 7 always means the same thing, on any review.

9–10 Exceptional Rare. Excels on this dimension with no meaningful caveat.

7–8 Recommendable Solid, with caveats worth knowing before you commit.

5–6 Fine, but outclassed Works, but a comparable alternative does this dimension better.

< 5 Avoid A real weakness on this dimension — not a nitpick.

two kinds of review, one rubric

Folium publishes two review types, always clearly labelled, and both are scored against the exact same rubric above:

Technical assessment: An objective, spec-level evaluation grounded in published specs, documentation, and independent sources — researched by an AI agent using web search, then reviewed before publishing. No ownership is claimed, and none is implied.

Hands-on review: Sylvaris's own money, bought at retail — no loan units, no early access, no embargo deals — and lived with for at least eight weeks before a score is published. This is the brand promise: bought, not borrowed.

about AI-drafted assessments

Technical assessments may be AI-drafted: an agent researches the product via web search, scores it against the rubric above, and drafts the review — but every claim must be grounded in what the research actually returned, with sources recorded, and a human can review before anything reaches the public site. The AI never manages affiliate relationships and never decides which products get covered — that stays a human decision (see our candidate list). If the available evidence is too thin to score fairly, the review is skipped rather than guessed.

revisited publicly

Scores are not permanent. When a firmware update, pricing change, or anything else shifts our assessment, the score changes — and the history of every change is shown publicly on the review itself. A score you can't question isn't a score you should trust.