The working, shown

Method

Three layers, in strict order

1 · Scripture first. If a name appears in a Sahasranāma we hold, that citation leads the page — list, scripture, position, teacher. This is the strongest evidence on the site.

2 · The Sanskrit engine. Dictionary attestation (Monier-Williams, Apte, Wilson), then grammatical resolution: every link between the spelling you confirmed and a dictionary stem is typed — same word inflected, a sandhi surface, a true derivative with shifted meaning, or an attested short form. A link we cannot type is a link we refuse to draw.

3 · Hindi, honestly labelled. Only when Sanskrit is silent do we consult Hindi sources, and the answer says “Hindi, not Sanskrit” in so many words — with the Sanskrit parent traced when Wiktionary records one.

Verified means verified

Where we claim a derivation, our engine has re-derived the word under Pāṇinian rules and the output matched your name exactly. Every sūtra number shown is read from that derivation’s own trail — never typed in by hand. Compound analyses that cannot be forward-derived are labelled a tier lower, as verified splits.

What our badges promise

  • Scripturally attested — the name is in a Sahasranāma we hold.
  • Verified derivation — a full derivation reproduced the name exactly.
  • Dictionary-attested — the whole word is a dictionary headword.
  • Verified split — the members are verified and re-join exactly; one tier below a full derivation.
  • Related form — a true derivative; both meanings shown, never merged.
  • Hindi, not Sanskrit — exactly what it says.
  • Partial — our search budget ran out with candidates unexamined. We tell you how many.
  • Not found — only when the search truly completed. A timeout is never reported as “not found”: a resource limit is not a fact about Sanskrit.

The wait is the proof

A hard name can take up to 45 seconds. The progress you watch is the real pipeline reporting its own stages — we would rather be slow and provable than instant and invented.