SNEWPapers

SNEWPapers gives Claude a research corpus built from the Library of Congress' Chronicling America newspaper collection. Every page is cut into components, a news story, an obituary, a runaway advertisement, a letter to the editor, a poem and many more types, each carrying the fully extracted text, a summary, the people and places it names, a genre label with structured fields for that type, a text-quality tier, and a link to the page on snewpapers.com and at the Library of Congress.

Claude follows a research method rather than guessing: it checks what the corpus covers for a place and a period (the index grows by tens of thousands of components a day), counts what matches a genre, a phrase or a topic before spending anything, reads pages of 25 summaries found by keyword, exact phrase or meaning, and opens only the components that matter, with their full text on request. Everything is read-only, and every result cites the paper, the date and the page with a link.

Historical newspapers print the past as it was. Anything with a caution flag may contain content offensive to modern readers; it is withheld unless the researcher asks for it. Nothing from a conversation reaches SNEWPapers except the requests Claude makes.

A SNEWPapers subscription or trial is required. Each subscription comes with a monthly token allowance; every call reports its cost and the remaining balance, and a free guide explains the whole method to Claude before it starts.