Paste an email's full source — headers optional — and get a SpamAssassin-style heuristic score with every triggered rule explained.
A heuristic approximation, not SpamAssassin itself. Scores are indicative, not verdicts.
No — it's a heuristic engine inspired by SpamAssassin's approach: ~30 weighted rules over headers, subject, URLs and body text, with the classic 5.0 spam threshold. Real SpamAssassin adds Bayesian filtering, network tests (DNSBLs, URIBL) and sender reputation, which a browser page can't replicate. Treat the score as indicative.
No. Paste the whole raw source for the full 30-rule run, or just the body/subject — header-dependent rules are skipped and the score is computed from what's there. In Gmail: open the message → ⋮ → "Show original" → copy all.
Under 5.0 looks clean, 5.0–10.0 is suspicious, over 10.0 is strongly spam-like. Legit marketing mail often lands between 2 and 5 — the triggered rules tell you which habits to fix, not just the number.
Common culprits: ALL-CAPS subject lines, URL shorteners, image-only HTML, missing List-Unsubscribe, and link text that doesn't match the href domain. Each rule explains itself — fix the hits, not the score.
Nowhere. All parsing and scoring runs in JavaScript in your browser. There is no server call at all on this page.