About Pundora’s corpus

Where the judgments and the law in Pundora come from, and what Pundora does with them.

The judgments

Judgments of every Indian High Court and of the Supreme Court, 1950 to today, held in Pundora’s own corpus with their full text. They are loaded from the courts’ own public archives, kept word for word, and topped up every night as the courts publish more. Every citation Pundora gives you is checked against the text of one of them, and every quotation it shows you is found in that text word for word or is not shown. No citation in Pundora is written by AI: the judgment is found first, and the citation is read off its record.

The law

Bare acts, section by section and word for word from India Code; the Schedule to the Limitation Act; the official tables that map the old criminal codes to the new ones; and the rules of civil procedure by Order and Rule. A section quoted in a draft is checked against these, not against memory. State legislation and regulators’ instruments come from the open-india-law dataset, used under its open licence with its attribution kept on every record.

Licensed sources

A small number of judgments and the links between judgments (which cites which) are obtained through the Indian Kanoon API under a commercial licence and kept in Pundora’s corpus with Indian Kanoon’s written permission. When Pundora looks something up there it sends only a citation, or search terms built from named legal parts. It never sends your documents, your own words, or a party’s name you typed.

Some judgments and citation links in Pundora’s corpus are obtained through the Indian Kanoon API under licence.

Your files are not the corpus

Your matters, your documents and your drafts live in your own encrypted, isolated workspace on India-region servers. They never join the shared corpus, and no AI is trained on them.