Docling parses the table. You need a workbook.
Docling gives you structure — cells, spans, reading order — but no .xlsx. Upload the same PDF here and get the spreadsheet step done: one sheet per table, numbers that stay numeric, and a source page behind every value so the output can be checked instead of trusted.
First file free, up to 10 pages · paid packs run up to 600 pages per file · files shredded after conversion
Analysts inheriting a parse
Someone (or some script) already ran Docling. You need the tables in Excel without re-running the model stack on your laptop.
Finance & accounting teams
Statements, schedules and filings where a single misread digit changes the answer — and where the reviewer will ask where the number came from.
Researchers & engineers
You use Docling in a pipeline but want a quick human-checkable export for the tables that matter, without writing an exporter per project.
What this actually fixes
Docling has no XLSX exporter
The pipeline emits Markdown, JSON, DocTags, HTML and chunks. Turning TableItems into sheets is a script you write and then maintain.
Markdown flattens what Excel needs
Merged cells, two-row headers and accounting negatives survive a data frame and die in Markdown — so the easy route is the lossy one.
Multi-page tables arrive as fragments
A statement that runs five pages comes back as five tables with repeated headers. Joining them, and keeping the page reference, is manual work.
What comes out
One upload, one set of compute-ready outputs — not a screenshot of a table.
One sheet per table
Each detected table becomes its own sheet with resolved header rows, spans filled, and values written as native numbers — not text that looks like a number.
Running-balance tie-out
For statement-shaped tables, every row is checked against the running balance so a misread digit or a dropped transaction surfaces before you build on it.
Click to the source page
Every value keeps its page and region, so an odd figure can be traced back to the PDF in seconds rather than re-keyed on faith.
Batch without mixing
Drop several PDFs at once; each file keeps its own sheet set and source references, with the file name carried into the output.
XLSX, CSV or Markdown
Export the format the next step needs — workbook for review, CSV for imports and scripts, Markdown when a human is reading.
Optional verification link
A tamper-evident record of which document produced which output, free for a third party to check — useful when the numbers go to a lender, auditor or client.
Documents we read
Native PDF, scan, phone photo, Word, Excel, PowerPoint, CSV, HTML, email or image.
How it works
Upload the PDF
Same file you would hand to Docling. Scanned and digital pages are handled in one pass; one file is one credit, up to 600 pages.
Check the table against the page
The parsed table is shown next to the source page. Confirm the header row, the signs and the row count before exporting — this is the step that makes a spreadsheet trustworthy.
Export XLSX or CSV
Download the workbook, or take CSV if a script or a system import is next. Values arrive as numbers, so SUM() works on the first try.
Every plan, every audience
- One file = one credit, up to 600 pages on a paid pack
- No merged cells; numbers arrive as numbers
- Every extracted value traces back to its source page
- Files shredded after conversion, never used for training
- A fresh file allowance every billing cycle; cancel anytime
- An optional hash-bound certificate anyone can verify for free
Questions
Does Docling export to Excel directly?+
Not natively — Docling emits Markdown, JSON, DocTags, HTML and chunks. Getting tables into .xlsx means writing an exporter (for example export_to_dataframe on each table) or using a converter like this one, which produces the workbook for you.
How is this different from running Docling myself?+
It is the same family of pipeline — Docling plus TableFormer table structure — with the spreadsheet step, the source-page preview and the running-balance check built in. You skip the environment; you keep the ability to verify each value.
What if I want to script it myself?+
That is a legitimate choice. Our guide at doclingocr.com/guide/docling-to-excel walks through export_to_dataframe, the JSON route for cell-level provenance, and the validation checks — including the four that catch most broken exports.
Will merged cells and two-row headers survive?+
A spreadsheet cannot represent merged cells, so spans are filled into the cells they cover and header rows are resolved into a single usable header. That is what makes the result calculable; the original layout is still visible on the source page.
How much does it cost?+
Per file, not per seat: $79 for 10 files, $199 for 30, $499 for 100, each up to 600 pages. The first file each day is free up to 10 pages — enough to check the output on your own documents before paying anything.
Are my documents stored?+
No. Files are shredded after conversion and are not used for training. If you need the conversion to stay entirely inside your network, run Docling locally instead — the guide explains how.
Income statement, balance sheet, cash flow and segments into Excel.
The statement-shaped case, with the running balance checked row by row.
When a script or an import needs rows instead of a workbook.
For reading rather than calculating — structure preserved, no spreadsheet.
Run your next file
Upload the document, get compute-ready output with the source page behind every number. The first file each day is free.