Our review methodology

We are an editorially independent directory of 4,921 AI tools across 26 categories. Advertising and sponsorship never change a ranking, a rating or the wording of a listing. This page documents exactly how listings are built, ranked and corrected, so you can judge how much weight to give anything you read here.

How a listing is built

1. Intake

A listing arrives from our imported dataset or a reader submission. We check the site resolves, the product is live rather than a waitlist, and the primary use case is stated clearly.

2. Deduplication

Products imported under several slugs are merged into one canonical listing, keeping the highest verified vote, view and rating values and the most complete description. Our last pass merged 45 duplicate groups and removed 66 redundant records.

3. Classification

Over a thousand tag-like source categories are mapped into 26 practical groups based on the job the tool does. The original category stays on the listing as a finer sub-label.

4. Verification

Pricing is only labelled when we could read a published plan on the tool's own pricing page. Logos fall back to the tool's own domain icon rather than a placeholder brand mark.

5. Ranking

Shortlists combine reader votes, review ratings and directory page views, then get a manual topical-fit pass. Signals with no data are excluded rather than counted as zero.

6. Maintenance

Dead-link sweeps, pricing spot-checks and duplicate detection run with every data refresh. Reader corrections go to the front of the queue.

What we publish

  • Pricing tiers we could read on the vendor's own pricing page
  • Ratings and votes collected from readers on this site
  • Derived pros, cons and plan tables, clearly labelled as derived from listing signals
  • Editorial articles with a named author and a stated testing approach
  • Paid placements, labelled as paid wherever they appear

What we never publish

  • Invented testimonials, credentials or case-study results
  • Ratings, review counts or benchmark numbers we did not collect ourselves
  • Imported pricing labels we could not verify — we hide them instead
  • Guarantees or performance claims a vendor cannot back
  • Ranking changes bought through advertising or sponsorship

How we handle missing data

Most imported records were incomplete, and an empty field is not the same as a bad score. A tool with no reviews shows no star rating instead of an empty zero-star row. A tool with no confirmed plan shows no pricing badge — that is a statement about our confidence, not about the vendor. Where a logo is missing we use the tool's own domain icon, and detail pages render a live screenshot of the site so you can see what you are clicking into.

Corrections and conflicts of interest

Listings go stale, categories get misassigned and prices change. If something here is wrong, tell us and we will fix the record or remove the listing. Paid placements are always labelled, and no vendor sees a ranking before it is published.