How we score

Every tool gets a score out of 100 from six weighted factors. The formula is public and the breakdown is shown on every listing.

Sponsorship has zero weight. Paid slots are labelled and kept apart from rankings. You can't buy rank here.

score = capability(30) + community(20) + value(15) + reliability(15) + momentum(10) + user_ratings(10)

Capability fit

30 pts

Tested against a standard task set for the use case by an agent, spot-checked by a human.

Community signal

20 pts

GitHub stars and growth, Hugging Face downloads, Product Hunt, X and HN mentions. Log-normalised and time-decayed.

Value

15 pts

Free tier and price for the core task. Open-source and generous free tiers score higher.

Momentum

10 pts

Releases and changelog activity in the last 90 days.

Reliability & trust

15 pts

Uptime, who is behind it, privacy and data terms, open source.

User ratings

10 pts

On-site votes, protected against brigading.

Freshness

Scores are recalculated weekly. Every listing shows a last verified date. Dead tools are archived automatically. Ties go to editorial judgement, and we note the reason.

Status of v1 scores

At launch, scores are editorial baselines set by our editors using the weights above. The automated test harness and community metrics are rolling out, and scores will update as they come in. Pricing marked “unverified” has not been checked against the vendor's pricing page yet.