Each tool page records published pricing, plan structure, core capabilities, notable limits, fit-score profiles, editorial pros and cons, and the verification timestamps tied to that record.

Comparison, best-list, use-case, pricing, alternatives, calculator, and preset pages are derived from that tracked layer rather than written as disconnected marketing copy, so the same core facts can flow into multiple buying paths.

Published pricing, plan inclusions, product positioning, and admin claims are checked against official vendor pages first. Pricing tables, product pages, help-center documentation, changelogs, and security pages are the default evidence sources.

The review workflow starts with the official pricing page, then checks the official documentation or help center for plan gates, usage limits, and admin prerequisites. If a rollout or governance claim matters, the supporting security or admin documentation must hold up as well.

When a buying claim matters, the review loop cross-checks the vendor's own pricing or product language across the rest of the public product surface so obvious contradictions are not carried forward as if they were settled facts. Third-party summaries can help with discovery, but they are not treated as verification.

Fit scores use a 1-10 scale across seven use-case dimensions: coding, research, meeting, automation, writing, customer service, and video generation. The score is not a benchmark trophy. It is a context signal showing how defensible the tool is for that kind of work.

A 10 means the tool is one of the strongest options we would defend for that workflow and buyer lens today. A middle score means the fit is real but caveat-heavy, and a low score means the workflow may still be possible but is hard to justify once tradeoffs, missing depth, or pricing friction are included.

Each tool is scored through three buyer lenses: individual, team, and enterprise. AgentHub does not collapse those into a single universal weight. The point is to show when a tool stays strong across rollout sizes and when the answer changes because governance, collaboration, or seat economics change the buying context.

The underlying dimensions include workflow breadth, execution depth in the native surface, rollout overhead, admin and governance posture, and price efficiency. For coding tools, that can include IDE depth and agent support. For broader assistants, it can include workflow coverage, handoff friction, and whether the published pricing still makes the recommendation easy to defend.

A higher rank means the tool is easier to justify for the stated workflow once fit, rollout overhead, pricing shape, and likely tradeoffs are weighed together.

The #1 tool is not treated as a universal winner. Category pages and use-case pages are deliberately scoped so the right answer can change with buyer context.

AgentHub distinguishes between free tiers, published self-serve paid starting prices, and quote-only deployment paths. Starting price language reflects the lowest comparable paid entry point that a buyer can actually act on from public pricing, not the most flattering marketing number.

Seat-based costs are interpreted with the stated billing model in mind. If pricing depends on enterprise quotes, credits, hidden usage limits, or plan prerequisites, the page treats the number as directional and surfaces that caveat instead of pretending the comparison is cleaner than it is.

We log verification dates and meaningful changes separately so pages can show both when the facts were last checked and when the buyer-facing meaning actually moved. A page can be freshly verified even if the recommendation did not materially change.

Published vendor pricing and product claims are time-sensitive, so outbound pricing links, product sources, and recent-change notes stay in the review loop. This is why pages can surface both `lastVerifiedAt` and `lastMeaningfulChange` style signals in different contexts.

Core tool records are reviewed at least weekly so pricing tables, plan prerequisites, and shortlist language do not drift for long stretches without a fresh pass.

When a material pricing, packaging, or access change is detected on a covered tool, the goal is to reflect that change within 48 hours if it could alter cost expectations, rollout assumptions, or shortlist position. Minor copy changes do not get the same urgency unless they change the buying meaning.

Affiliate relationships, sponsorship, and outbound attribution do not change fit scores or ranking order. If a tool monetizes better but is a weaker workflow match, the weaker editorial position stays in place.

Commercial infrastructure sits downstream from the editorial call. The operating rule is simple: workflow match, rollout overhead, pricing interpretation, and documented tradeoffs decide the recommendation, not revenue potential.

AgentHub does not publish filler rankings, fake review counts, invented benchmark wins, or generic verdicts written only to capture traffic. If the evidence layer is too thin, the safer choice is to leave the scope narrow or leave the gap visible.

Commercial routes do not buy a better rank. Affiliate or partner plumbing can coexist with a recommendation page, but it does not override workflow match, pricing interpretation, or documented tradeoffs when the shortlist is ordered.