Research OS is built for accounting, finance, economics, and management PhDs chasing JAR, TAR, JF, RFS, and JFE. It reads your paper library, remembers the decisions you made about your sample and identification strategy, and drafts at the rigor bar of your target journal — not a one-click paper generator.
Not a generic AI wrapper. Each skill is tuned for social science empirical workflows — WRDS data conventions, identification checklists, journal-specific style, honest citation flags. Your decisions carry across sessions so you never re-explain your sample.
Daily digests from SSRN, NBER, and arXiv tuned to your research agenda — the sources social scientists actually read, not just ML preprints.
Reads the papers you upload. Drafts systematic, narrative, scoping, meta-analysis, or annotated reviews in the voice of your target journal, with honest [VERIFIED / UNVERIFIED] citation flags.
Understands Compustat, CRSP, IBES, ExecuComp, Audit Analytics. Walks you through DiD, PSM, IV, RDD identification checklists before it writes any code. Exports reproducible Stata .do and Python with clustered SE, winsorization at 1/99, and a standard robustness menu.
Drafts one section at a time in JAR/TAR/JF/RFS/JFE style. Remembers the research question, sample, and identification strategy you locked in across prior sessions. Flags [YOUR CONTRIBUTION] and [CAUSAL CLAIM] so you never accidentally ship AI-fabricated results.
Before long-form output, Research OS asks about your research question, sample, identification strategy, and expected direction. You can always skip the questions — but the defaults favor rigor over speed, because your dissertation committee and journal reviewers will.
Upload your reading list. Every skill has the full text in context — not a Google Scholar snippet.
Citations cross-checked against Crossref. Anything unverified ships with a visible warning badge, in the draft and in the exported .tex.
Confirm your sample period once. Research OS remembers it across every session, in every skill, until you unlock it.
Not inflated user counts. Concrete commitments you can verify on day one.
Start free with 50 credits/month. One credit ≈ one AI skill run. A typical empirical paper uses 15–25 credits across literature, design, testing, and drafting.
Research OS is in closed beta for social science PhDs. Request access and start with the four skills that understand your workflow.