No vision model, no rasterising
The DXF is read programmatically from its entity data. Nothing is rendered to an image and no picture is sent to a vision model, so text is transcribed exactly as stored rather than recognised from pixels.
DXF to Excel
A DXF already contains its text as data — title block entries, notes, dimension values, revision rows and block attributes are all addressable entities. Reading it as an image would throw that away, so AxioExtract parses the file itself and hands you the rows.
Parsed, not looked at
Schema mapping only
A lightweight model is used for one job: mapping the parsed entities onto the output columns. That keeps the read cheap and deterministic where it matters.
Charged per file — 6 credits per file
CAD files are priced per file rather than per page, because there are no pages to rasterise.
What you get back
The title block, general notes, dimension values, revision history, and block or balloon attributes read from the drawing's own text layer. Because these come from the file rather than from character recognition, part numbers, thread callouts and tolerances arrive as written.
Rows are reviewed in the grid and exported as an Excel workbook (.xlsx), tab-separated values for Google Sheets, JSON, or XML. Full lists are on supported formats.
Formats that need converting first
DWG, Revit (.rvt), SolidWorks, Inventor, CATIA and STEP/IGES are not read directly. Export to DXF for 2D drawings, to PDF for issued sheets, or to IFC for building models — see IFC to Excel. For drawings issued as PDF, the engineering drawing profiles handle the read instead.
Common questions
How do I get data out of a DXF file into Excel?
Upload the DXF with the CAD drawing profile. The file's entities are parsed programmatically rather than read as a picture, and the resulting title block, notes, dimensions, revisions and block attributes can be reviewed and exported to Excel, Google Sheets, JSON or XML.
Is the DXF read by an AI vision model?
No. DXF files carry their text and attributes as data, so they are parsed directly and no image is sent to a vision model. A lightweight model is used only to map the parsed entities onto the output schema.
Can DWG files be uploaded?
No. DWG is not supported. Export the drawing to DXF, or to PDF, and upload that instead.
Try it on your own document
Credits are included with every new workspace — no card required to start.
Start extracting