UnPlotter is a browser-based tool for recovering numerical data from figures in PDF files — the inverse of plotting.

The Problem

Technical papers, reports, and product data sheets routinely publish data only as figures. Fortunately, these documents are usually published as PDF files with high quality figures included in a vector format. When you need the numbers represented by the curves — to validate a model, compare with your own results, or reproduce an analysis — you’re left tracing curves by hand or taking a screenshot and using a raster plot digitizing tool. UnPlotter automates this with an accurate, calibrated workflow.

How It Works

  1. Load a PDF directly in the browser (nothing is uploaded; processing is local)
  2. Define the axis calibration by selecting curves representing x and y extent
  3. Select and name the curves you want to export
  4. Export as CSV or JSON

Axes can be linear or logarithmic. Extract multiple curves per figure with just a few clicks.

Implementation

UnPlotter runs entirely in the browser using JavaScript and the PDF.js library for rendering and parsing. No server, no data upload, no account required.