Sources and licences

The Sanskrit of the Bhagavad Gita is an ancient work. It is in the public domain everywhere. This app follows the traditional recited text in Devanagari. Chapter 13 keeps Arjuna’s opening questions as verse 1, so that chapter has 35 verses and the whole text has 701. People often call the book “the 700,” depending on how one verse is joined or dropped. Nothing is omitted here.

The Roman transliteration is a mechanical IAST rendering of that public-domain text, with sandhi left as it is sung and word-splits added for study. Word meanings, the Hindi translation, the English translation, and both commentaries were written for this app. They are not quotations. They are informed by the commentarial habits of Ramanuja, Madhva, Vallabha, Sridhara Swami, and the Chaitanya tradition, and they say so only where the verse supports the point. If those traditions differ, the note says so briefly. They do not copy any modern book.

Not included, by design: no text from the Bhaktivedanta Book Trust, no translation or commentary by A.C. Bhaktivedanta Swami Prabhupada, and no other copyrighted modern translation or commentary.

The speaking voice is synthetic, not a recording of any person. Male or female reads the whole verse, in this order: the Sanskrit, each word once and then its Hindi meaning, the Hindi translation, the Hindi commentary, the English translation, and the English commentary. A bare name takes its case in the meaning, such as ने, को, से or का. Sanskrit and Hindi use hi-IN. English uses the accent chosen in Settings, Indian English (en-IN) or US English (en-US). Only chapters that already have saved audio are played. A chapter that is not recorded yet is marked audio coming soon and is not sent to the voice service.

Talk to Krishna is a language model. It is labeled in the app as AI guidance inspired by the Gita, not a real person. It is not a guru, a doctor, a lawyer, or a political adviser.