Historical research,rebuilt for the AI era.
PrimarySourceFinder is an AI research assistant I built to gather, verify, and analyze primary and secondary sources in minutes, so you can spend your time understanding history instead of hunting for it.
My mission
Primary sources are the foundation of honest research, but they live scattered across thousands of archives, libraries, journals, and museums. Most of them are free. Almost none of them are easy to find. I built PrimarySourceFinder to close that gap, and to give students, teachers, researchers, and history enthusiasts the same caliber of research environment a professional historian would use.
What the app does
Five capabilities, one workspace, one continuous research flow.
AI identifies real primary and secondary sources for any topic, ranks them by scholarly importance, and surfaces free copies across open-access archives, peer-reviewed journals, books, PDFs, articles, images, and YouTube.
Add any source to Sofia, the built-in AI research assistant. She reads the actual content of the page and answers questions about it. No more guessing what's behind a paywall or a 400-page PDF.
Fact-check claims against the live web, run a bias check, generate a deep dive, build a timeline, study the historiography, and pull key takeaways. Every source comes with the full analytical toolkit.
Save sources into a workspace and draft your essay in the Notetaker, a rich-text editor with 28 AI commands for writing, tone, style, and translation across 11 languages. Export to .docx when you're done.
Turn any source into a podcast, narrative story, quiz, or flashcards. A simplified mode rewrites dense material in plain English when a topic is brand new.
Why I built it
I am a history enthusiast, not a credentialed historian. But I understand what primary sources are and why they matter, and I kept noticing the same problem everywhere I looked online. People repeat historical stories without ever pointing to where the information came from. Threads, articles, videos, even textbooks. Claims float around with nothing underneath them.
The other half of the problem is that when I do go looking for the primary source behind a story, finding it is hard. Archives are scattered, badly indexed, or buried so deep you give up before you reach them.
I love studying history. I built PrimarySourceFinder because I was tired of the gap between caring about evidence and being able to actually find it.
How it is different
Search engines find pages. Chatbots invent citations. PrimarySourceFinder does neither.
Built for history enthusiasts, not generalists
Sources are evaluated for scholarly importance instead of SEO rank, with priority given to primary documents from academic archives, government records, and museum collections so the right material rises to the top.
Open-access first
Domain boosts favor archive.org, .gov, .edu, Wikimedia, ResearchGate, and Semantic Scholar. The goal is to put real sources in your hands, not paywalls.
From question to finished document
Most tools stop at search results. PrimarySourceFinder carries you through reading, fact-checking, analyzing, and writing in one workspace, so the work doesn't fragment across five different tabs.
Verifiable by design
Sofia cites the actual page content she's reading. Fact-check cross-references claims against fresh web search. You can always trace a statement back to the source.
Start your first investigation.
New accounts get 200 free credits. Enough to research a paper, fact-check a source, and draft an essay before you ever reach for your wallet.