Editorial rules · updated 2026-09-11

Independent scores need visible rules.

This is how AIToolsEssentials evaluates tools, handles uncertainty, checks pricing, uses benchmarks, and keeps monetization separate from editorial judgment.

Open evidence ledgerRequest a correction

01 · What the score means

Scores measure decision fitness, not popularity.

Our editorial score is a structured judgment about how well a tool fits its stated job for a defined buyer. A high score does not mean the tool is best for everyone. Every review includes best-fit guidance, trade-offs, and a test-before-paying checklist.

Job fit · 30%

Does the tool solve the workflow it claims to solve, at the buyer's likely scale?

Output quality · 25%

How useful, controllable, and repeatable are the outputs for the target job?

Workflow depth · 20%

Does it support the surrounding process — files, integrations, agents, export, or collaboration?

Value & cost pressure · 15%

Does the pricing model make sense at realistic usage, including limits and seat pressure?

Trust & operational fit · 10%

Are privacy, rights, availability, and policy details clear enough for the intended buyer?

02 · Source hierarchy

Official sources outrank summaries.

  1. Primary: vendor pricing, documentation, terms, privacy, rights, and product announcements.
  2. Secondary: reputable benchmark publishers and clearly identified third-party evidence.
  3. Context: community reports, labeled as individual experience rather than editorial testing.
  4. Unknown: missing information stays marked unresolved; we do not fill gaps with confident-sounding guesses.

See the source row for every reviewed tool →

03 · Freshness rules

Fast-changing claims get dates.

Pricing

Official pricing rows show the last checked date. Pricing Watch compares future snapshots and records confirmed differences rather than inventing history.

Models

Comparisons name actual shipping model/product versions and checked dates. A vendor lineup change triggers a data update before editorial copy is treated as current.

Benchmarks

Benchmark entries retain the exact model, benchmark name, configuration, result, and snapshot date. Missing comparable data is labeled unavailable.

04 · Monetization independence

Affiliate links cannot buy a ranking.

Affiliate and sponsored links are labeled and use the required sponsored/nofollow attributes. We only wire tracking URLs issued by a real program. Affiliate availability, commission, or conversion rate does not change an editorial score or placement.

Where no approved affiliate relationship exists, the page uses the vendor's ordinary official link. Grammarly and Impact-linked programs remain excluded until an approved tracking relationship exists.

05 · Community evidence

Experience reports stay labeled.

We do not fabricate hands-on results, votes, traffic, or testimonials. Community reports include the tool, plan, task, duration, what worked, what broke, and cost context. Published reports remain attributed as individual experiences and do not silently become editorial test results.

Challenge the record

Found a stale claim?

Send the URL, the relevant source, and what changed. Corrections are logged and source data is rechecked in the next maintenance run.

Submit a correctionCheck pricing freshness