Enlightened is a conversation with a patient teacher of the world's sacred texts, with receipts: real passages, exact quotations, and a verification check on every quote. The category's open problem is that AI misquotes scripture; Enlightened's answer is a teacher who cannot get away with it.
No. The engine retrieves real passages from public-domain translations and the model may quote only what is in front of it. Every quotation in every answer is machine-checked against the source text, and the verification result is shown under the answer.
Enlightened says so plainly and describes what the retrieved passages do contain. It never fills gaps from memory. Honest insufficiency is a designed feature, not an error.
A corpus built on the Internet Sacred Text Archive's public-domain translations: the King James Version and World English Bible, the JPS 1917 Tanakh, Rodkinson's Talmud, Pickthall's and Palmer's Quran, the Bhagavad Gita, Upanishads and Mahabharata, the Dhammapada and core suttas, Legge's Texts of Taoism and Confucian classics, the Zend-Avesta, Jaina Sutras, the Kojiki, the Guru Granth Sahib in English, and more. Translation and era are part of each citation.
No. It is a study tool. It describes what texts say with citations; it does not issue rulings, prescribe practice, or speak for any tradition. Questions that need a human authority are pointed to one.
A conversation with a teacher. You can follow up, ask for another tradition's view, or go deeper on one line; the teacher remembers the thread and ends each answer by offering where you might go next. Underneath, every turn runs a fresh search of the texts, so the conversation stays grounded.
Reading and searching the texts is free forever. Each message to the teacher costs one credit; everyone gets free credits every day, and credit packs are available for longer conversations. Your conversation lives in your browser, not on our servers. Questions are never sold, never used for advertising, never used to train models, and never written to server logs. Sending a message to a model provider happens only after explicit consent.
Ask your first question