Data Mint

Your documents become rows, your questions become columns. Every cell linked to evidence.

A platform for research-grade data extraction using privacy-first AI.

Developed by Karl Rohe, Professor of Statistics at UW-Madison

Request Access See the demo Log In

Click any cell to inspect the evidence and reasoning

This is what your minted data looks like. Try it.

Systematic Review & Evidence Synthesis
Extract PICO elements, risk of bias assessments, and outcomes with complete citations. Every cell traced to source.
Evidence & Reasoning
📋
Click any cell above to see detailed evidence and reasoning

Three components make Data Mint
systematic, automatic, and transparent

1 · Codebook

Makes it systematic

Specifies exactly what to extract from each document.

  • Describe your project — bring an existing protocol, sample documents, or just your goals
  • A guided conversation structures your expertise into a codebook
  • Human-readable, easy to edit — defines variables, decision criteria, and edge cases
2 · Minting

Makes it automatic

Multiple AI models extract independently, then discuss to verify.

  • Discussion verifies evidence and reasoning
  • Replicate extractions measure inter-rater reliability
  • Complete audit trail for every cell
3 · Inspection

Makes it transparent

Click any cell to view the evidence, quotes, and reasoning behind each value.

  • Full provenance trail for every extraction
  • Confidence levels highlight uncertain results
  • Refine codebook and re-mint in minutes

Inspect → refine → re-mint. Forgot a field? Add it. Found an edge case? Update the codebook. Minutes, not weeks.

Built for research rigor

10 min
First dataset
10,000
Documents per batch
100%
Cells with provenance

PDFs, images, Word docs, HTML, and more. Any file containing text or visual information. Systematic reviews, qualitative coding, evidence synthesis — and beyond. Predictable pricing — a penny per page, a penny per cell.

Built in Madison, Wisconsin

Karl Rohe · Professor of Statistics, UW-Madison · Founder

Craft your codebook, mint the data.

LinkedIn

Ready to mint your first dataset?

Request Access Log In

Frequently asked questions

How fast is it?

Most users mint their first dataset in under 10 minutes. Once your codebook is ready, batch processing scales to thousands of documents automatically.

How do I verify results?

Click any cell to view the exact source text, key quotes, and the model's reasoning. Every extraction includes complete provenance.

What if I need more fields?

Add new fields to your codebook and re-extract in minutes. Iteration is built in. Codebooks are reusable and shareable.