Engineers do not keep their plots as tidy PNGs. They live inside PDF datasheets — a MOSFET's safe-operating-area curve on page 4, an op-amp's gain-bandwidth plot buried on page 9. Every other digitizer makes you screenshot the figure first, crop it, save it, then upload that image.
Plot Extractor opens the PDF itself. It renders the pages, shows thumbnails, and detects the plot figures on each one. You point at the chart you want and start extracting — the screenshot detour is gone. It is the first free digitizer built to work on datasheet PDFs directly.
Not all PDFs are the same, and the difference decides how accurate an extraction can be.
A vector PDF stores a plotted curve as geometry: the actual path — line segments and Bézier curves — that the line follows. The data is, in a real sense, still in the file. A raster plot is just a grid of pixels: a picture of a curve, with no structure underneath.
| Vector PDF | Raster image / scan | |
|---|---|---|
| Curve is stored as | Geometric paths | Coloured pixels |
| Best you can recover | The exact drawn path | A trace of the pixels |
| Typical accuracy | Near-perfect | Very good, limited by resolution |
Plot Extractor uses whichever is available. When a PDF page carries real vector graphics, it reads the path geometry directly — that is as close to the original data as digitizing can get, with no pixel guesswork. When the page is a scan or a flattened image, it falls back to its pixel-tracing engine. You get the most accurate method each plot can support, automatically.
Reading vector graphics out of a PDF is genuinely difficult. The format was designed to print a page, not to hand back structured data. Curves arrive as streams of drawing operations with their own coordinate transforms; gridlines, axis frames, and text are mixed in with the data and have to be told apart. Most digitizers skip all of it and treat every PDF as a flat picture.
Doing the harder thing is the point: a curve recovered from vector geometry is not an approximation of the plot — it is the plot.
There is one more reason this matters, and it is not about accuracy.
Plot Extractor runs entirely in your browser. PDF parsing, OCR of the axis labels, figure detection, curve tracing — all of it executes locally on your machine. There is:
For a hobby chart this is merely convenient. For an engineer working from a confidential, pre-release, or NDA-bound datasheet, it is the whole point — the document physically cannot leak through a tool that never transmits it. Privacy here is not a policy promise you have to trust; it is a property of how the tool is built.
Open a PDF, pull out the curve, export the numbers — and the file stays exactly where it started.