AdiVeda publishes 324 devotional texts across eleven categories — chalisas, aartis, stotras, ashtakams, kavachams, mantras, namavalis, the Durga Saptashati, the Sundarkanda and others. Each is presented in Devanagari with IAST transliteration, a phonetic reading line, and Hindi and English meaning.
Sacred text carries an obligation that ordinary web content does not. A single wrong syllable changes a mantra. This page sets out exactly where our texts come from, what we check, and — just as importantly — what we have not checked.
Where the texts come from
Every text page names its source in the Text source & editorial review panel at the foot of the page. Depending on the text, that is the composing poet or seer (Goswami Tulsidas, Adi Shankaracharya, Budha Kaushika), the scripture it belongs to (Markandeya Purana, Rigveda, Ramcharitmanas, Mahabharata), or the published repository the Sanskrit was taken from.
Where we relied on a public repository, we say so and name it. The repositories we have drawn on and checked against include sanskritdocuments.org, GreenMesg, StotraNidhi, BhaktiBharat, DrikPanchang and SriMandir, alongside commonly circulated print editions.
How we check them
Our review happens at two levels, and we tell you which one applies to the text you are reading.
Read verse by verse. The text was read line by line against one or more published editions. The Devanagari, the IAST and the phonetic line were each checked. Where the live text disagreed with every edition we could find, we corrected it. This applies to 125 texts — every chalisa, every aarti, every kavacham, and the most-recited stotras, ashtakams and mantras.
Checked for script integrity and completeness. The text was scanned for the specific defects we know creep into Devanagari on the web, and confirmed to end with its proper closing colophon — but it was not read line by line against a published edition. This applies to 183 texts. We label these plainly rather than let them borrow the credibility of the deeper pass.
A further 15 pages are navigation hubs that carry no scripture text of their own; we verified that every text they link to exists and resolves. One text, the Vishnu Sahasranamam, is a separate build in progress and carries no review date at all, because claiming one would be false.
What the script scan looks for
These are not hypothetical faults. Each was found in our own corpus and fixed.
क्श written for क्ष. The dominant error in the chalisa collection, affecting words like परीक्षा, लक्ष्मी and क्षण. Corrected in Ganesh, Shani, Ram and Bhairav chalisas among others.
Zero-width joiner artifacts. An invisible character (U+200D) inserted after श, breaking शृ in words such as शृणु and शृङ्ग, and leaking into the transliteration. Found in four kavachams and six Durga Saptashati texts, all removed.
Avagraha leaking into transliteration. The Devanagari mark ऽ appearing untranslated in the IAST and phonetic fields. Four occurrences found corpus-wide and fixed; IAST now renders it as an apostrophe, as in dharmo ’dhikatamo.
Missing framing verses. Dropped opening shlokas and closing dohas. The Jai Ambe Gauri aarti was missing five concluding verses; Om Jai Jagdish Hare was missing its final verse. Both restored.
Non-authentic texts. Twice we found a wrong composition published under the right name — the Baglamukhi Chalisa, and the Kunj Bihari aarti, whose fourth verse described a sword and trident that belong to no authentic Krishna aarti. Both were rebuilt from verified sources.
Where we deliberately leave things alone
A received reading that faithfully reproduces a commonly published edition is not an error, even when a purer form exists. We preserve सहस्त्रांशु, the Govindashtakam’s योशिद्, and the common व/ब interchange in बभूबुः, because these are what the tradition actually circulates. Correcting them would mean silently substituting our judgement for the printed tradition.
Where a reading looks wrong but we cannot confirm the correction, we flag it and change nothing. Fifteen texts currently carry such an open question, marked on the page itself. We would rather show you an unresolved doubt than publish a confident guess.
Name-count variation in the ashtottara-shatanamavali sets is likewise left alone. Published recensions genuinely differ on how many names they carry; a list of 107 or 110 that numbers itself consistently is a variant, not a mistake.
What we do not claim
We have not consulted manuscripts. We have not carried out academic or peer-reviewed textual criticism. We have not established a critical edition, and we do not adjudicate between recensions.
We do not claim the corpus is error-free. A text marked as read verse by verse was compared against published editions by careful human-directed review, not by scholarly apparatus. A text marked as script-checked may still contain a reading we have never examined.
What we do claim is narrower and, we hope, more useful: that we tell you which of those two things was done, when it was done, and against what.
Found something wrong?
Corrections are genuinely welcome, particularly from readers who recite these texts daily or hold a print edition we have not seen. If you can name the edition you are reading from, we can act on it quickly. Please get in touch with the text name, the verse, and the reading you believe is correct.