Organic ranking

Principles

We never seed or invent votes, ratings, clicks, or views. Organic order is computed from how recently the free offer was checked, free-offer value, optional editorial picks, real listing dates when we have them, and recorded community activity. Paid placements are labelled and sit outside this order: payment can't buy rank.

For a short overview and our disclosure, see /how-we-rank. This page is the full reference the support chat uses.

The formula (high level)

Each tool gets an organic core score from weighted signals. Signals are log-scaled and compared to peers in the same category (or site-wide when a category has fewer than 20 tools). The final rank score is:

rank score = freshness (F) × free value (M) × organic core

Free value M = 0.70 + 0.30 × FV, where FV blends how long the free offer lasts, its usability (counted only when at least two checks are known) and whether “no card” is confirmed.

Trending is separate: it only activates after at least 5 trusted events in the last 90 days, then blends velocity and recent activity with freshness.

Freshness (F)

Freshness reflects how recently we verified the free tier on the public pricing page:

• 1.0 when verified within 7 days
• 0.92 when verified within 14 days
• 0.75 when verified within 30 days
• 0.5 when verification is older
• 0.5 when not yet verified
• 0.3 when the pricing link appears dead
• 0 when the free tier is marked gone

Signals (E, R, U, C, V, X, N)

• Editorial (E) — Curator picks from catalog data; “tool of the week” comes from these picks.
• Ratings (R) — Bayesian average with prior mean 3.5 and prior weight 10 when data is sparse.
• Upvotes (U) — One per voter (a voter ID cookie, or a salted IP hash without one), with a 30-day half-life.
• Try-free clicks (C) — Smoothed CTR; smoothing alpha 200.
• Velocity (V) — Last 72 hours vs a 28-day baseline.
• External (X) — Optional Product Hunt / GitHub-style fields; capped at 0.1 of the core so it cannot dominate.
• Newness (N) — From the date a tool was listed on the site; decay scale 10 days. Every tool in the catalog at launch (19 Oct 2026, 1:30 AM IST) counts as listed at launch, so none gets a head start. We never guess listing dates.

Base weights (before renormalization)

Weights shift as real events arrive. At low confidence (few trusted events), editorial and newness weigh more. At full confidence (50+ trusted events in 90 days), editorial and newness weigh less.

• Editorial (E): 0.4 → 0.10000000000000003
• Ratings (R): 0.15
• Upvotes (U): 0.15
• Clicks (C): 0.1
• Velocity (V): 0.1
• External (X): 0.1
• Newness (N): 0.15 → 0

Only signals with data for a tool are included; weights are renormalized over present signals.

Anti-gaming caps

• No single non-external signal carries more than 35% of the weight; the rest goes to the other signals (with only one or two signals they share it equally).
• External (X) contribution is further limited to 0.1 of the combined core.

Ask why a tool ranks where it does

Use reader support chat or compare live breakdowns — the site uses the same function for ordering and for explanations.