ClinCoder · TLF programming

Photo-to-Program

Photograph a table shell, confirm what was read, and get an R program and a SAS program back, with the R already run and every number set beside the photo.

Request a demo Open the live studio Earlier version

Upload, paste or photograph a table shell, the analysis spec page and an old program for the same table. Nothing is written until a person has checked the reading and the derivation: the tool has 7 stages and 2 human gates before any program exists.

What you get

An R program, executedRun on the server against the study's analysis datasets, or against bundled synthetic data until you upload yours.
A SAS program, written, not runpendingThere is no SAS on the server. The SAS is labelled pending until an optional kit runs it on your own PC and compares cell for cell.
The output tableRTF and PDF of the table the R produced.
Shell against outputThe photographed table and the produced table side by side; every cell that differs is listed. It says that they differ, not which side is wrong.
A gap reportRows it could not derive print PLACEHOLDER and are named, never filled with a guess.
An audit trail and download bundleEach confirmation, correction, run and sign-off is appended to a hash-chained log that ships in the zip.

How it fits together

Photo-to-Program flow: two human gates sit between reading and generation. Shaded amber steps are people; blue steps are the software. illustrative

Screens

Synthetic demo run on the live v2 studio, 2026-10-30: the shell as read (top) and the table the R program produced (bottom). Rows the demo could not derive print PLACEHOLDER and are listed as disagreements.

Same synthetic demo run: the generated SAS, marked as not yet executed in SAS, and the optional PC verification kit.

The in-app help states what the tool will not pretend: SAS is written, never run here; demo data mismatches any real study.

What it does not do yet

SAS is generated but not executed on the server; it stays labelled pending until the PC kit reports back.
Until you upload your own analysis datasets the R runs on invented demo data, so a real study's photo will always disagree with it.
A sign-off is an application record, not a qualified electronic signature.
The demo screenshots were taken from a run on synthetic data; they show the tool reporting a mismatch, which is what it should do on that data.

Talk to us

Demos run on synthetic data. Write to bhanoji.d@gmail.com with the table or study type you want to see.