<!-- Generated from the public HTML page by tools/generate-discovery.mjs. -->

Source: https://www.quantumsteeldesign.com/pdf-rescue/

Free local drawing recovery

# Convert eligible structural steel PDF drawings to DXF.

Recover vector linework and semantic text from technical PDFs as an editable DXF.
Files stay in your browser—there is no drawing upload, account, or cloud conversion fee.

[Convert in your browser](https://www.quantumsteeldesign.com/pdf-rescue/#browser-converter)
[Download Desktop + CLI](https://www.quantumsteeldesign.com/pdf-rescue/#desktop-download)

- Runs locally

- Vector PDF recovery

- DXF R12 browser output

- Detailed recovery report

PDFengineering drawing

LOCAL

DXFrecoverable geometry

Your drawing never crosses this browser boundary.

Public tool counters

## Useful enough to keep around?

Loading anonymous totals. No drawing names or drawing contents are collected.

↓
—
Desktop downloads

DXF
—
Browser conversions

Browser edition

## Choose a steel drawing PDF and recover eligible vector content.

Local processing · 150 MB maximum · modern desktop browser recommended

Recovery complete

## Your files are ready.

Exploded lettering or unusable geometry?
Quantum Steel Design can rebuild drawings, models, shop details, and revision-ready steel documents.

[Request professional help](https://www.quantumsteeldesign.com/contact/?source=pdf-rescue-browser)

Windows desktop edition

## No installer. No uninstall. Extract it and run it.

Quantum Steel PDF Rescue is a self-contained, portable Windows x64 application. It does not add a
Windows service, startup item, or traditional installed program entry.

- Download the ZIP and choose Extract All. Do not run the program from inside the ZIP.

- Open the extracted app folder.

- Double-click PdfToDxf.App.exe to open the desktop app.

To remove it later, close the program and delete the extracted folder—there is no uninstaller to run.
The desktop edition adds PDFium extraction, R14/R2000 output, CAD cleanup, native curves, repeated blocks,
experimental dimensions, drag-and-drop batches, detailed logs, and explicit existing-file choices.

[Download Windows x64](https://downloads.quantumsteeldesign.com/Quantum-Steel-PDF-Rescue-0.2.0-beta.3-win-x64.zip)
[SHA-256 checksum](https://downloads.quantumsteeldesign.com/Quantum-Steel-PDF-Rescue-0.2.0-beta.3-win-x64.zip.sha256)
[CLI documentation](https://www.quantumsteeldesign.com/pdf-rescue/cli/)

Desktop v0.2.0-beta.3 ·
released Aug 3, 2026, 7:30 AM EDT

Unsigned beta · portable · no installer · no separate .NET installation · verify the published checksum before opening

CLI INCLUDED

### Automate repeat conversions.

The same download includes cli\PdfToDxf.Cli.exe for scripts, batch jobs, and repeatable conversion settings.

.\cli\PdfToDxf.Cli.exe convert drawing.pdf `
--output .\DXF `
--dxf-version r2000 `
--keep-both
Run .\cli\PdfToDxf.Cli.exe --help for every option.

[Open the complete CLI guide](https://www.quantumsteeldesign.com/pdf-rescue/cli/)

Honest recovery

## A structural steel PDF drawing is not the original CAD model.

### Vector paths

Lines and curves can usually be recovered as LINE or POLYLINE geometry. Browser fills become conservative closed boundaries.

### Lettering

Semantic PDF text can become DXF TEXT. Lettering stored as tiny outline paths remains exploded geometry because the original characters are gone.

### Raster scans

Photographs and scanned sheets are reported but not vectorized. They require redraw, OCR, or a separate tracing workflow.

Use a small, checkable sample first

## From a PDF page to a reviewed DXF

Conversion creates candidate geometry. A useful result also needs the correct page, units, scale, and receiving-application checks.

### 1. Identify what the PDF actually contains

Choose a page with recognizable linework and a known reference dimension. Text that can be selected in a PDF viewer may be recoverable as text; lettering drawn as outlines is still geometry. A scanned sheet may look clear at normal zoom but contain no recoverable vector paths. A document can also mix vector content and raster images.

### 2. Establish units and scale

The browser converts PDF page coordinates into the selected output units and applies the drawing-scale multiplier. A multiplier of 1 preserves the page's physical size in those units. It does not automatically infer the real-world scale of a plotted drawing.

For an illustrative 1:48 drawing, a 1-inch line on the PDF page represents 48 inches, so a scale multiplier of 48 is the starting value. Verify a second known dimension after importing the DXF. Enlarged details, reduced plots, and pages containing several drawing scales need separate attention; one page-wide multiplier cannot resolve different scales within a sheet.

### 3. Read the recovery report

Keep the JSON report with the exported DXF. It records the selected pages, units, scale, flattening tolerance, per-page results, and reported limitations. Counts of lines or text entities show what was recovered; they do not prove that every visible object was recovered correctly.

Review image and clipping warnings. A raster image is reported rather than traced into geometry. PDF clipping or paint effects can make recovered candidate paths differ from the visible boundary, so inspect those areas against the source page.

### 4. Inspect the result in the receiving CAD application

Check a known dimension in each relevant direction, the page orientation, line connectivity, lettering, duplicate geometry, and important curves. Keep the source PDF visible during that review. Use the result for its reviewed purpose and retain the controlling drawing issue alongside it.

Interpret the result

## Common recovery problems and the next useful check

### The size is wrong

Check the output units, scale multiplier, original plot scale, and receiving application's import-unit setting. Measure known source dimensions instead of changing the multiplier until the drawing merely looks plausible.

### The lettering is hundreds of small lines

The PDF may contain outlined glyphs rather than semantic characters. Those paths cannot be turned back into reliable editable wording by assuming their appearance. Keep the outlines as reference or use a separately reviewed text-reconstruction workflow.

### There is no usable geometry

Confirm whether the page is a raster scan, whether the intended page was selected, and whether its content is within the converter's documented support. OCR can help recover text from a scan; it is not the same as rebuilding accurate CAD geometry.

### Curves look segmented

The browser emits conservative R12 LINE, POLYLINE, and TEXT entities. Curve tolerance controls approximation after the selected output transformation; smaller values can increase the number of segments and file size. Compare the curve to the source rather than treating a denser polyline as proof of accuracy.

Keep drawing comparison and geometry recovery separate: [compare two PDF issues](https://www.quantumsteeldesign.com/drawing-compare/) to review visual changes, or [search selected PDFs](https://www.quantumsteeldesign.com/drawing-search/) when the task is locating a note or reference. Neither step requires converting the document to DXF.

When recovery is only the first step

## Turn reference geometry into steel documents the shop can use.

[Start a project](https://www.quantumsteeldesign.com/contact/?source=pdf-rescue-bottom)
[Detailing services](https://www.quantumsteeldesign.com/services/)

A few practical answers

## Frequently asked questions

Why does a scanned drawing produce little or no geometry?

A scan may contain only pixels. PDF Rescue recovers suitable vector paths and semantic text; it does not infer reliable CAD geometry or dimensions from a raster image.

How do I check the DXF scale?

Compare a known labeled dimension against the recovered geometry in your CAD application, confirm units, and inspect another dimension in a different direction. Page size alone does not establish model scale.

Is the original PDF uploaded?

The browser conversion runs locally. Keep the original PDF and compare the exported result before using it in another workflow. The separate anonymous usage-statistics choice is optional.
