Reviewed evidence · deep regional context · follow-up questions

Ask the Museum.

Bring your hardest history question. Explore Mercer County and its neighbors, Civil War campaigns and statehood, coalfields and railroads, Black history, communities, genealogy, architecture, landscapes, and the records historians use to test a claim.

How answers are controlled →
A history buff's research companion

Ask broadly. Follow the evidence deeply.

The agent searches every indexed public MCHS page, all published collection APIs, and a reviewed source layer spanning local and regional history. It cites what it used, suggests the next question, corrects common spelling variants, and says when the evidence is insufficient.

What “trained” means here

Wide coverage, narrow claims.

The agent is designed for discovery without pretending every local-history question is settled.

Complete site index

Every public page and section in the sitemap is rebuilt into the searchable corpus during validation.

Reviewed source layer

Institutional, government, university, and MCHS guidance supplies high-value context beyond the current collection database.

Fail closed

If reviewed evidence does not support an answer, the agent identifies the gap and points to the research desk instead of guessing.

Research boundaries

  • Questions stay in the browser unless you explicitly enable optional Azure AI synthesis when it is available.
  • Only public, published records are loaded. Private applications, inquiries, registrations, and unpublished visitor notes are excluded.
  • A generated or extractive answer is a research lead. Open its cited records before repeating a claim.
  • Living people, burial records, traumatic history, property access, and culturally sensitive material require additional care.