Canonical text first
AI may assist search, presentation, transliteration review, or explanation, but it must not silently rewrite sacred source text.
About SlokaSandra
Sacred text deserves care, context, and visible provenance.
SlokaSandra is a devotional reference from Slightbook designed to help readers discover, read, understand, and learn Indian slokas, mantras, stotras, hymns, and related traditional texts without hiding where the material came from or how it was reviewed.
Why we are building it
A devotional library should preserve the prayer before it explains the prayer.
Devotional texts often circulate across books, temples, families, languages, recitation traditions, transliteration systems, and online copies. Small differences can matter. SlokaSandra is being built to make those differences visible rather than silently normalizing them into one AI-generated version.
Our goal is to keep the original script, language, transliteration, meaning or translation, source work or traditional attribution where verifiable, deity and tradition relationships, editorial review status, and audio provenance together in one reference model. When a source is uncertain or traditions differ, the page should say so plainly.
Provenance stays visible
Source work, chapter or verse, traditional attribution, editorial status, and verification notes should remain attached to the content where available.
Tradition-aware context
Meanings, recitation practices, names, and interpretations can differ across sampradayas, regions, languages, teachers, and editions. Those differences should be respected.
Audio is treated as evidence
Human-verified recitation, AI-assisted pronunciation, and unverified audio are distinct states. Pronunciation-sensitive material should not be labeled authoritative without review.
Part of Slightbook
Built as a long-term devotional reference product.
SlokaSandra is a Slightbook product. Canonical-text corrections, stronger source references, transliteration concerns, pronunciation feedback, accessibility issues, partnership enquiries, and product suggestions can be sent through the contact page.
Reader feedback is especially useful when an edition differs from the displayed text, a traditional attribution needs qualification, a translation is too narrow, or audio should carry a different verification status.