From a spreadsheet of headers to an evidence-linked table

EviSearch reads your column headers (and one example row), writes a definition for every column, extracts each value from clinical-trial papers with the page it came from, and flags the cells a person should check. Every correction can become a rule the next paper benefits from.

    Papers in the demo table
    —
    Reported values
    —
    Flagged for review
    —
    Rules learned from reviews
    —

    Try it in three minutes

    1. Open the demo table: 133 columns for metastatic hormone-sensitive prostate cancer trials, drafted from the column headers and one example row, then reviewed and locked.
    2. Review a flagged cell: see what Agent A (reads the whole paper) and Agent B (searches it) answered, what the Reconciliation Agent decided and the page it checked; confirm or correct it.
    3. Turn a correction into a rule and see which columns it would change; then open Learning to see what your reviews changed.

    Everyone shares one workspace in this demo: reviews and rules you save are visible to the next visitor.