Methodology · कार्यपद्धति

How this site is made

This page says plainly where each kind of content comes from, what is machine-assisted, and what has not been reviewed by a person. It describes what is true today, not what is intended.

1. Source texts

The Sanskrit text is imported from open digital archives, mainly sanskritdocuments.org and the DharmicData collection on GitHub. Each chapter file records the archive address, its licence and the date it was fetched, and the verse page shows the archive it came from.

These are digitised texts, not critical editions prepared by this site. They have not been compared line by line with a printed edition, so spelling, sandhi and verse numbering can differ from the edition you use. Numbering follows the imported source.

2. Translations

Where the source archive supplies a Hindi or English translation, it is used as supplied. Where it does not, a translation was drafted with AI assistance. Every such verse carries an “AI” flag in its data, and the reader marks it on the page.

As of 8 October 2026, 72,471 of 187,405 verses in the library carry that flag. Those translations have not been checked by a Sanskrit scholar. Treat them as reading aids, and check the Sanskrit and a printed translation for anything that matters.

3. Commentary and attribution

Classical commentators (for example Shankara, Ramanuja and Sridhara) are named only when a summary is attributed to them. Those summaries are short paraphrases written for this site. They are not quotations or translations of the original commentaries, and they have not yet been checked against the original works. Commentary that was produced with AI assistance is labelled as such.

4. Modern explanations

Simple meanings, “why it matters today”, modern examples, misunderstanding notes, teaching flows and reflection questions are editorial. They were written for this site to help new readers, and they are labelled “not scripture” wherever they appear. They are drafts awaiting scholarly review. They make no promises about outcomes and are not medical or psychological advice.

5. AI assistance

AI (Anthropic’s Claude, among other tools) has been used to draft Hindi and English translations, commentary summaries and the editorial explanations above, and to help build this site. AI can make errors, including confident-sounding ones. That is why such content is flagged, kept visually separate from the Sanskrit, and open to correction.

6. Human review

No systematic human review of the texts, translations or commentary is recorded yet. For that reason the site does not show “verified” or “reviewed” badges. When a named reviewer checks a text, the review will be recorded with the reviewer’s name and the date, and only then will a badge appear.

7. Corrections

Every verse page has a “Report correction” button. It opens a prefilled issue on GitHub. Nothing is sent until you press “Create issue” there, and this site stores nothing about you. Fixes are made in the project repository, whose public history is the record of what changed and when. There is no separate correction log on the site yet.

8. Rights information

The classical Sanskrit texts are ancient and not under copyright, but a particular edition, translation or commentary can be. The source library names the edition behind each text. Where the licence of an edition or translation has not been confirmed, that is not stated as settled here; if you hold rights in material shown on this site, use the correction link on any verse page to tell the editors and it will be reviewed. Explanations marked as editorial or AI-assisted are this project’s own drafts.

9. Version history

The site has no separate version log. Every change to texts, translations and explanations is recorded, with its date, in the public project repository history, which is the authoritative record. Pages that carry an editorial explanation show its last-updated date where one is recorded.

Back to the scripture library.