Leaderboard

The cleanest pSEO sites on record.

Lower is safer. Ranked by SpamBrain risk score, one entry per domain. Methodology and category breakdown below the table.

Current rankings

1
paperforge.dev preview

paperforge.dev

Generate professional legal and business document templates for free. Fill a form, preview live, and download your PDF instantly.

A5
2
tonfoyer.fr preview

tonfoyer.fr

L'argent que tu envoies à ta famille au pays est déductible de tes impôts en France.

A7
3
pseolint.dev preview

pseolint.dev

Audit your programmatic SEO site for SpamBrain risk in 60 seconds.

A14
4
astrostella.fr preview

astrostella.fr

Horoscopes, tarot, chiromancie et thèmes astraux propulsés par l'IA. Découvre Stella, ton guide céleste.

D60
5
bestfirenze.com preview

bestfirenze.com

Discover the best places to eat, visit, stay, and shop in Florence, Italy. Curated recommendations from locals and travelers for an authentic Florentine experience.

D60
GradesA0–19B20–39C40–59D60–79F80+Lower = safer

Leaderboard methodology in one paragraph: the ranking is rebuilt every 10 minutes from completed public audits, deduplicated by hostname so a domain occupies exactly one slot, and sorted ascending by composite score with createdAt as the tiebreaker. Audits below the 5-page floor are excluded because too-small samples produce volatile rankings. Pages marked private by their owner never appear, regardless of score. Listings inherit the retention window of the underlying audit, so anonymous entries fall off after 24 hours and free-tier entries after 30 days unless the operator upgrades and pins them. The board first shipped on March 15, 2026 alongside the v0.4.0 engine cut, and the scoring weights were last rebalanced on April 21, 2026 when the AEO category landed.

The pseolint leaderboard ranks programmatic SEO sites by their composite risk score — a 0-to-100 number where lower is better. The score is a weighted aggregate of findings across SpamBrain + AEO rules weighted by your site's archetype (programmatic-directory, blog, ecommerce, docs, small-marketing). Each finding contributes severity- and confidence-weighted points; the site-type profile decides which rule families dominate the final score.

The dominant signal is the spam category, weighted at 33%, because doorway-style scaled abuse is what Google’s SpamBrain classifier targets most aggressively on programmatic sites. Content quality contributes 19%, AEO readiness 14%, internal links 11%, and the rest (tech, data, schema, cannibal) divide the remaining 22%. A site that ships clean templates with real answers can finish well below 30 even at scale; a thin doorway farm will rarely score below 60.

We publish the leaderboard so operators can benchmark against peers without paying for opaque enterprise dashboards. It also demonstrates that the engine produces stable, repeatable results: the same site audited twice in a week should land within a few points. If your audit tells a different story than your traffic, that’s a signal worth investigating — usually a templating regression or a recently shipped doorway pattern.

Ranges to keep in mind: A = 0–19 (ready), B = 20–39 (good), C = 40–59 (concerning), D = 60–79 (severe), F = 80+ (critical). Anything in the A band is healthy enough that we wouldn’t expect manual-action exposure on the SpamBrain axis. C-band sites are where most undermaintained pSEO domains live; D and F is where deindexation events tend to start. The thresholds were calibrated against the August 2022 helpful-content rollout and the March 2024 core update, so a site that lands in the A band today should also have been clean against those historical baselines.

How sites end up on this leaderboard

Any audit a user runs with isPublic = trueshows up here once it completes and crosses the 5-page minimum. Free-tier audits cost $0 and default to public — that’s the trade for unlimited one-shot acquisition runs, capped at 3 audits per browser per 24-hour window. Pro plans start at $19/mo, default to private, and stay private unless an operator explicitly flips the visibility toggle on the audit detail page.

Listings are deduplicated by hostname; only the most recent audit per domain appears, and rankings refresh every ten minutes. Audits expire after their retention window, at which point they fall off the board automatically. If you ran a public audit you didn’t mean to share, you can mark it private from your dashboard and it disappears from the next revalidation.

Scoring methodology

Audits crawl up to 200 pages on the free tier and 500 on Pro (manual re-audits), sampling URLs from the sitemap and the homepage’s outbound links. Render mode is opt-in via the --render flag — useful for SPA frameworks that hydrate content client-side, but skipped by default to keep audits fast and deterministic. Each sampled page runs through every rule in the engine, and findings are bucketed by severity.

The score is computed by summing severity weights (critical = 40, error = 25, warning = 12, info = 5) per category, capping each category at 100, then multiplying by the fixed category weight and summing. The 8 spam/* rules and 8 aeo/* rules are the biggest individual contributors because they map directly to the patterns search and answer engines penalise. A clean, well-templated site can run hundreds of pages and still land in the A or B band; a site that trips even one critical doorway rule across many pages will jump into the C/D band quickly.

  • spam33%

    Doorway pages, scaled abuse, thin-content templates — the patterns SpamBrain targets directly.

  • content19%

    Coverage breadth, helpful-content depth, evidence of expertise rather than filler boilerplate.

  • aeo14%

    Answer-engine readiness — answer-first paragraphs, citable facts, modular self-contained sections.

  • links11%

    Internal linking sanity, anchor diversity, no orphaned pages or stuffed footer link farms.

  • tech7%

    Crawlability, status codes, render-blocking assets, sitemap and robots hygiene.

  • data6%

    Data-source freshness and citation — does the page back up its claims with verifiable inputs?

  • schema5%

    Structured data correctness — valid JSON-LD, no spammy or misleading markup.

  • cannibal5%

    Keyword cannibalisation — multiple pages competing for the same query without differentiation.