Skip to content

AI research for scholarly books

AI that actually
reads your sources.

Ask a research question. Four specialist agents pull real data, trace theoretical lineages, dig up historical precedents, and poke holes in the argument — all at once, with citation verification built in.

Built for the humanities and social sciences. Free during early access.

husearch

Research Query

Data

Waiting…

Theory

Waiting…

History

Waiting…

Critic

Waiting…

Synthesized Research Brief

Key Finding: Both Japan and the US deployed QE as unconventional monetary policy, but divergent outcomes…

Empirical Evidence

Japan’s QE (2001–2006, expanded 2013) expanded the BOJ balance sheet by 350%, yet inflation…

Counter-Arguments

Critics note that QE disproportionately inflated asset prices, widening wealth inequality…

Sound familiar?

You annotate a PDF in one app, write in another, manage citations in a third, and lose the thread between all of them.

You submit a draft, wait three months for reviewer comments, and find out you missed a major counter-argument on page two.

Your co-author is editing v7_final_REAL.docx while you’re on v7_final_REAL_edits2.docx.

The workflow

You ask once. Four agents go to work.

01

Ask your question

Plain language, any scope. “How did QE outcomes diverge between Japan and the US?” works just fine.

02

Agents dig in parallel

Data, Theory, History, and Critic agents each search real sources — FRED, Semantic Scholar, JSTOR, arXiv — simultaneously.

03

Read a research brief

A synthesized write-up with cited claims, counter-arguments flagged, and a bibliography you can actually use.

The four specialists

Data

Pulls GDP, inflation, employment numbers from FRED, World Bank, IMF, and more.

Theory

Searches Semantic Scholar, OpenAlex, CrossRef, and arXiv for relevant frameworks and debates.

History

Digs through Internet Archive, Library of Congress, and academic databases for precedents.

Critic

Looks for what contradicts your thesis — methodological gaps, rival findings, things you might miss.

The workspace

Everything in one place.

Research, annotate, write, and collaborate without switching tabs. Your sources, notes, and drafts stay connected.

Research output

Briefs and bibliographies, automatically

Every research session gives you a structured brief with an executive summary and an annotated bibliography — tagged by chapter, searchable later. An agent log shows you exactly how the findings were gathered.

BriefBibliographyAgent Log

Key Finding: Divergent QE outcomes…

Empirical Evidence: BOJ expanded…

Policy Implications: Central banks…

Annotation

Read and annotate in the same place

Highlight a passage, attach a note, run source criticism on a PDF — provenance, reliability, bias — without switching apps.

Bernanke_2015_QE_Analysis.pdf

The effectiveness of large-scale asset purchases depends critically on monetary transmission mechanisms which operate through the portfolio rebalancing effect.

Under liquidity trap conditions, conventional policy tools lose traction.

Compare with Friedman’s channel theory

Peer review

Catch the gaps before your reviewers do

Run a peer-review simulation on any chapter. It flags weak arguments, missing evidence, and methodological concerns so you can fix them before submission.

AI Peer Review
3 comments
Argument

The causal link between QE and employment needs stronger evidence. Consider adding the Blinder (2010) counterfactual analysis.

Methodology

Selection bias concern: Japan and US cases aren’t directly comparable without controlling for demographic differences.

Sourcesresolved

Strong use of primary BOJ transcripts. Consider adding Eurozone ECB data as a third comparison point.

Collaboration

Stop emailing drafts

Co-authors and advisors edit the same document at the same time. Real-time cursors, threaded comments, role-based permissions. One version of the truth.

Y
SP
CO
3 editing
Live

The divergence in QE outcomes between Japan and the United States reveals fundamental differences in how monetary transmission operates across institutional contexts.

Bernanke’s approach prioritized signaling effects, while the BOJ focused on direct balance sheet expansion.

This distinction matters because

SP

Dr. Patel: Should we cite Koo (2008) on balance sheet recessions here?

You: Good call — adding it now.

Writing

Write where you research

Footnotes, citations, track changes — a real manuscript editor. No more copying findings into Word and fixing the formatting for an hour.

BIH₂“”fn

The Japanese experience with quantitative easing provides a crucial counterfactual for evaluating the Federal Reserve’s response to the 2008 crisis.

1 Bernanke argues that the scale of intervention…

Argument mapping

Map your argument visually

See how your claims, evidence, and counter-arguments actually connect. Spot the gaps in the structure, not just the prose.

ThesisClaim AClaim BEvidenceEvidenceCounterEvidence

A research assistant who never tires, and a librarian who has read everything you cite.

You do the thinking. Husearch does the legwork.

This isn’t a chatbot that writes your paper. It’s a research tool that goes and finds what you’d find yourself — the data, the precedents, the counter-arguments — just faster. Every judgment call is still yours.

You set the direction

You decide what to investigate, which sources matter, and when to stop.

You check the sources

Every claim links to a real paper or dataset. Nothing is taken on faith.

You write the paper

The brief is raw material. The argument, the framing, the voice — that’s yours.

FAQ

Questions, answered.

How is this different from asking ChatGPT to do my research?

The agents call live academic and data APIs — Semantic Scholar, OpenAlex, CrossRef, arXiv, FRED, the World Bank, Internet Archive, the Library of Congress — so findings trace back to real papers and datasets, not a model’s recollection. A dedicated Critic agent actively hunts for counter-evidence, and the synthesizer flags contradictions rather than smoothing them over.

Where do citations come from, and how do you avoid fabricated sources?

Every cited claim links to a record returned by a real source API, and references are verified against CrossRef in the background. When a source can’t be resolved it’s surfaced, never invented — and an agent log shows you exactly which databases each agent searched.

Can I import my Zotero library?

Yes. Husearch reads CSL-JSON, BibTeX, and RIS, so a Zotero (or Mendeley, or EndNote) export drops straight into your citation manager and is reusable across every chapter.

Who owns my drafts and notes? Is my work used to train models?

Your work is yours. We don’t train models on your projects, drafts, or notes. Everything is stored in your account, organized by project and chapter, and exportable whenever you want it.

Can I export to LaTeX or Word for my publisher?

Yes — export a chapter or a full manuscript to .docx or print-ready PDF, or generate LaTeX for a journal template. Footnotes, citations, and formatting come along with it.

Does it work offline?

The workspace is a progressive web app, so your drafts and recent sessions stay readable and editable offline. New research runs — which call external APIs — pick back up once you’re reconnected.

What disciplines is it built for?

The humanities and social sciences. A per-project Research Focus tells the agents your discipline, regions, time period, key thinkers, and critical lens, so results are steered to your field instead of staying generic.

Start the book you’ve been putting off.

Early access. Built for researchers who’d rather think than search.