Skip to main content
Editorial

How we score AI tools

Every tool on ToolHund gets a single 0–100 ToolHund Score. It is not a popularity contest and it is not paid for. Here is exactly how it is calculated and what it does — and does not — measure.

The five signals

  • Review quality — 45 points. Weighted average of community ratings once a tool has cleared the review threshold. Below the threshold we use editorial baseline only.
  • Popularity — 25 points. Logarithmic scale of active users (public sources, vendor claims we could verify, or third-party traffic data). Log scale so a 200M user tool does not crush a great niche one.
  • Feature depth — 15 points. Number and breadth of core capabilities that materially change what the tool can do.
  • Value — 10 points. Pricing transparency and the presence of a usable free tier, freemium plan or open-source option.
  • Freshness & trust — 5 points. Editor's picks, recently updated tools and tools with a clear changelog earn the ceiling here.

Review threshold

A tool must have at least 5 community reviews before its aggregate community rating counts toward the review-quality signal or is exposed as structured data to search engines. Under the threshold, we clearly show "Editorial rating — awaiting community reviews" and disclose the sample size.

Freshness policy

Tool records show the date they were last verified by a human editor. Pricing that has not been re-checked in 90 days is flagged as "Verify latest". Broken links are removed within 7 days of the next crawl.

What the score does not measure

Sponsored placements

Tools can pay to appear in clearly labelled "Sponsored" or "Featured" slots. Those slots never change a tool's ToolHund Score, category ranking, or comparison outcome. See our full disclosure.

Corrections

If a score, price, or feature is wrong, tell us on the tool's page or email hello@toolhund.com. Verified corrections are applied within 48 hours.

About ToolHundEditorial policyDisclosure