SNEWPapers Research is a remote MCP server for researching historical US newspapers, 1736 to 1963, built from the Library of Congress's Chronicling America collection. Connect it to Claude, ChatGPT, Gemini, Grok, Cursor or any MCP client that supports remote servers with OAuth, then ask questions in plain language.
Every newspaper page is split into its individual labelled components, e.g. domestic news, editorial, poem, obituary, and 28 more. Each item is labelled, summarised, has label specific (and queryable) meta-data extracted, and is quality-graded and linked to the original Library of Congress page scan, so answers come with citations you can check. The index holds about 6.8 million searchable components as of November 2026 and grows daily by about 100k new ones.
For an example conversation demonstrating common features, look at https://snewpapers.com/blog/snewpapers-research-features
Tools (11, all read-only)
research_search: keyword, phrase, semantic and hybrid search with date, place, title and genre filters
research_facets: counts over the matching items by decade, state, title or label, and much more
research_component, research_components: full item text and scan links
research_corpus, research_titles, research_title: what is covered, by newspaper and period
research_labels, research_vocabulary, research_guide: the labels, the controlled terms, and how to research well
research_account_tokens: your balance
Access
Comments