About & testing methodology

ToolProofs is a small independent test lab that publishes hands-on comparisons of SaaS and AI tools. Every review is researched, tested, and written by our in-house team under the byline ToolProofs Test Lab. We don't use ghostwriters and we don't take vendor briefings — what you see is what we actually found during testing.

How we test

We use two evidence grades, and every page labels which one it used:

• Hands-on (tested) — we created a real account (free plan; we do not enter card details), performed scripted scenarios with Playwright/Chromium plus manual checks, and archived timestamped screenshots, exported files, and network payloads. Every hands-on claim links to a file in the page's evidence index. Our raw test log (pre-registered test plan and filled results) is kept for every hands-on page and sent on request — ask for a page's log fromContact & corrections.
• Terms check (documented) — we verified a claim against the vendor's official page on a stated capture date and screenshotted that page. These rows are labeled "official-page verified" or "docs only" and must not be read as first-hand experience.

Comparison tables mix the two grades deliberately, and every mixed table carries aTested by us? column so the asymmetry is visible per row, not buried in prose.

What we don't do

• We don't enter card details. Anything behind a paid wall is marked as a documented check, never as a test.
• We don't review a tool from its marketing page alone. If we could not capture a claim ourselves, the page says the claim is unverified — or drops it. When we can only check a vendor's own homepage or pricing page, the page carries a "homepage-level verification" label and is never presented as a test.
• We don't hide the parts that went badly. Bugs we hit, signup walls we could not pass, and limits we never got to exercise are listed as coverage gaps on the page.
• We don't let affiliate status change a verdict, a score, or the order of a table.

How a page gets made

A review starts as a written, pre-registered test plan — the scenarios and the expected result for each case are fixed before anything is run, so a case can't be quietly dropped later. Collection is scripted (Playwright/Chromium against a real free account, plus manual steps where a product fights automation). Drafts are then assembled from those captures, checked line by line against the evidence files and the vendor's own pages, and edited before publishing; nothing ships with a first-hand claim that has no capture behind it. Each page keeps its own evidence index, and the pre-registration logs behind them stay on this site as plain text — sent on request rather than listed publicly, through Contact & corrections.

Scores

Hands-on reviews score six dimensions (core features, onboarding, reliability, team fit, language support, pricing value) on a 0–5 scale. Scores are anchored to specific evidence files listed in the scorecard, and the overall rating is the arithmetic mean of the six — recomputed, never hand-set.

Keeping prices current

Vendor pricing changes constantly, so every page that quotes a price carries the date we last checked it against the vendor's own pricing page. Older pages are re-captured on a rolling basis, and when a number has moved we update the page, the price table, and that check date together rather than leaving a stale figure in place. Two details we keep explicit: which billing state a price belongs to (monthly and annual figures often differ by 25% or more, and we print which one we captured), and whether a "free" tier is a permanent plan or a trial.

Corrections

When a factual error is identified, we fix the live page, state the correction on the page itself, and date it (see the dated correction block on ourTally alternatives page). We do not silently rewrite history.

How we make money

Some links on this site may be affiliate links. When a page contains affiliate links, it says so at the top of that page. No tool on this site has paid for placement or a score today, and none can: reviews are produced first, published second, monetized (if at all) third.

Contact

Corrections, evidence requests, and testing-methodology questions:hello@toolproofs.com. We answer correction reports within a few days and credit the reporter on the page when the fix is material.