Copying a PDF table into Excel by hand works for one page. It stops working the moment the data lives in a title block, an exploded parts list, a P&ID or a 12-page invoice batch. This page covers the three practical routes, when each one is worth using, and how AxioExtract handles the documents the others struggle with.
Excel can import a PDF table directly via Data → Get Data → From File → From PDF. It's free and fine for a clean, text-based table. It fails on scanned pages, merged cells, rotated sheets and anything where the values are labels on a drawing rather than rows in a grid.
OCR turns pixels into text. That gets you words, not structure: you still have to decide which string is the drawing number, which is the revision, and which balloon number belongs to which part. On technical documents this is usually where the time goes.
Instead of reading text and hoping, you tell the tool what kind of document it is — engineering drawing, exploded view with BOM, P&ID, invoice, or an office table — and it returns the fields that document type is supposed to contain. That's how AxioExtract works, which is why the output columns are the same on every file.
Drop in single or multiple files — PDF, PNG, JPG, DOCX, XLSX or CSV — and pick the document profile that matches them. Up to 12 pages are read per file.
The page sits beside the results grid with each extracted value highlighted on the document, and low-confidence readings flagged. Correct any cell, or draw a box over a region to re-read it.
Download a formatted Excel workbook (.xlsx), export the same rows as JSON or XML for an ERP or API, or copy them straight into Google Sheets.
A 12-page supplier invoice PDF is charged at 1 credit per page, so 12 credits in total. An assembly drawing set of 3 sheets is charged at 2 credits per page, so 6 credits. Every new workspace starts with 50 credits included, which covers roughly four 12-page invoice batches before you buy anything.
Yes. Pages are rasterised and read as images, so scans and photographs of printed sheets work the same way as native PDFs.
Exploded views are read as a diagram plus its bill of materials, so item numbers are kept with the part number, description and quantity. You still review every row before exporting — that's the point of the review screen.
Excel (.xlsx), Google Sheets, JSON and XML. See supported formats for the full list of inputs and outputs.
They sit in a private bucket scoped to your workspace with row-level security, and you can delete the stored file from the server while keeping the history record.
50 credits are included with every new workspace — no card required.
Start extracting