Batches of PDFs, one column layout

PDF Scraper: Scrape Data From PDF to Excel With PDF Scraping Software

Drop in a stack of invoices, statements or monthly reports and get every table back as rows and columns, scanned pages included. The same scraper takes 20 files per batch in the browser or one file per call through the API, and it bills per file, so a 40 page statement costs the same as a one page receipt.

For finance, operations and data teams who pull the same tables out of the same kinds of PDF every week.

Uploading your PDF…

Drop your PDF here or click to browse

PDF files up to 50MB

Scrape one of your own files and check the rows before you pick a plan.

The short answer

A PDF scraper reads the tables inside PDF files and returns them as structured rows you can load into Excel, CSV or a database. It is not a web scraper: a web scraper collects the files, a PDF scraper reads them. For recurring business work, judge a scraper on three things: whether it keeps columns aligned across page breaks and scans, whether it handles many files in one run, and what unit it bills. On 200 twelve page statements a month, a per page plan costs about $269 and a per step AI platform over $800 before formatting, while a per file plan costs $24.

PDF scraping is two jobs, and most tools only do one

People searching for a PDF scraper usually have one of two problems, and the tools that rank for the phrase solve different halves of it. Getting clear on which half is yours saves a trial of the wrong product.

Half one

Collecting the PDFs

Crawling a portal, a supplier site or an inbox and saving every PDF it links to. Web scrapers, browser extensions and email rules do this well. None of them read what is inside the file.

Half two

Reading the tables inside

Turning each file into rows with the right columns, keeping page breaks from scrambling the grid, and reading scans. This is the hard half and the one PDFxlsx does.

If your PDFs already arrive by email or sit in a shared drive, you only have half two. If they live on a website, a small collector script plus our PDF to Excel API joins the halves with one POST per file. Microsoft 365 teams can wire the same thing without code using Power Automate PDF to Excel.

How much does PDF scraping software cost? Three real workloads priced

Scraping volume is described in documents, but most scrapers bill pages or workflow steps. The table prices the same three monthly workloads on each vendor's published list rates. Parseur figures come from its own pricing API, monthly billing. Nanonets figures use its Starter rate of $1 per credit and count only import, AI extraction and export, before any formatting charge. The full rate tables are in our breakdowns of Parseur pricing and Nanonets pricing.

Monthly workload Pages Parseur (per page) Nanonets (per step) PDFxlsx (per file)
1,000 one page invoices 1,000 $129 (Starter, 1,000 pages) About $380 (40 import + 300 extraction + 40 export credits) $49 (Pro, unlimited files)
200 twelve page bank or card statements 2,400 $269 (Premium, 3,000 pages) About $824 (8 + 720 + 96 credits) $24 (Plus, 500 files)
One 400 page annual or loan report 400 $129 (Starter, the 300 page plan is too small) About $136 (0.04 + 120 + 16 credits) $9 (Starter, 50 files)

Where the others come out ahead

Parseur bills an email or a CSV as one credit whatever its length, so data that arrives in the email body is cheap there. Nanonets offers volume discounts of up to 40% on its Growth plan and does approvals and ERP posting we do not. Annual billing lowers Parseur by roughly 20 to 26%.

Why the gap is widest on statements

Statements are long and table heavy. Every extra page adds to a per page meter, and on Nanonets every extra row adds formatting runs on top. A per file plan ignores both, which is why the middle row shows a gap of more than ten times.

What a PDF scraper has to get right when it runs on a whole folder

One file is a conversion. Two hundred files every month is a scrape, and the failure modes change. A mistake that you would spot on a single sheet slips through when the rows go straight into a pivot table. These are the checks worth running on any tool before you trust it with volume.

The same columns on every file

If March comes back with six columns and April with seven, the stack breaks. Consistent output per document type is what makes a batch mergeable.

Tables stitched across pages

A transaction table that runs over 12 pages should not come back with the header repeated as data 11 times or the carried forward balance counted twice. Test this on your longest file first.

Scans in the same run

Real folders mix digital and scanned files. A scraper that silently returns nothing for a scan is worse than one that fails loudly. Ours runs OCR on scanned PDFs automatically.

Numbers that sum

Amounts stored as text, currency symbols left in, and (125.00) instead of a negative number all break a SUM without an error message. Values should land as numbers.

A row count you can check

The API response reports pages, tables and rows found for each file. Compare it with the printed total before the rows flow anywhere else.

No per sender setup

Template scrapers need a new set of zones for each new layout. When you scrape PDFs from dozens of suppliers, setup time is the real cost.

How to scrape data from PDF files with PDFxlsx

1

Gather the files

Save the PDFs from the inbox, portal or drive into one folder. Group them by document type if you plan to stack the results.

2

Upload a batch

Drop up to 20 files at once on Plus or Pro. For a scheduled scrape, POST each file to the API instead.

3

Let it read every page

Tables are detected on every page, each with its own headers, and scanned pages go through OCR without a setting to switch on.

4

Export and stack

Download XLSX or PDF to CSV, or read JSON from the API, then append the rows into your master sheet or database.

The batch workflow in more detail, including how to name files so the merged sheet stays traceable, is on the batch PDF to Excel converter page.

PDF scraping tools compared by the job they fit

Stated plainly, including the jobs where another kind of tool is the better buy.

Tool type Examples Handles a folder Scans Fits best
Code libraries pdfplumber, Camelot, Tabula Yes, if you write the loop No, needs a separate OCR step Developers with clean digital PDFs and time to tune
Web scrapers and extensions Crawlers, browser extensions, hosted scraper actors Collects files well Mostly text only Half one: downloading the PDFs a site links to
Template scrapers Docparser, Parseur Yes, per template Yes A few senders with stable layouts, email-borne data
AI document platforms Nanonets, Docsumo, Astera ReportMiner Yes Yes Named fields routed through approvals into an ERP
PDFxlsx, table-first This page Yes, 20 per batch or via API Yes, automatic OCR Every row of every table, many layouts, billed per file. Not a fit for approval routing.

If you are choosing between these for a wider parsing project rather than a scrape, PDF parsing tools and software compares the same categories by output and billing unit. Coming from a library, the Tabula alternative page covers what changes when scans and borderless tables enter the folder.

What teams scrape out of PDFs every month

Supplier invoices

Line items from dozens of vendors, each with its own layout, stacked for spend analysis. See invoice PDF to Excel.

Price lists and catalogs

Distributor price sheets scraped each quarter and compared line by line. See price list PDF to Excel.

Long statements and reports

Brokerage statements, loan tapes and filings running to hundreds of pages. See large PDF to Excel.

Operational exports

Aging reports, inventory counts and system printouts that only leave the old system as PDF. The PDF table extractor shows the detection step alone.

PDF scraper questions

Upload the PDF to a PDF scraper that detects tables, let it read every page (running OCR on scanned ones), then download the rows as Excel or CSV. For more than a handful of files, upload them as a batch or send them to an API so every file comes back in the same column layout and can be stacked into one sheet.

The best PDF scraper is the one whose output and billing unit match your job. For every row of every table across many files, a table-first scraper billed per file is usually cheapest. For a few named fields from one stable layout, a template scraper such as Docparser or Parseur fits. For developers with clean digital PDFs, pdfplumber or Camelot costs nothing but engineering time.

Yes, but only with a scraper that runs OCR first. A scanned page is an image with no text layer, so a library that reads text returns an empty result rather than an error. Our scraper detects pages with no text layer and runs OCR on them automatically, so a folder that mixes digital and scanned files goes through one upload.

Use a scraper that processes files in a batch and keeps the same columns for the same kind of document. Upload up to 20 files at once on the Plus and Pro plans, download each result, and stack the sheets, or pull the rows as JSON through the API and append them in your own script. Check the row count per file before you merge.

A web scraper can download the PDF files a site links to, but it does not read the tables inside them. That is a second job. The usual pipeline is a crawler or a script that collects the PDF links, then a PDF scraper that turns each downloaded file into rows. Our API covers the second half, so the two steps connect with one POST per file.

It can read a short PDF and describe or retype a small table, but it is not a scraping pipeline. Long tables get summarized or truncated, output varies between runs, and there is no batch mode or row count to check against. For a table you need complete and identical every month, a dedicated scraper with a fixed output is the safer choice.

It depends on the unit. Parseur bills one credit per PDF page, from $49 a month for 100 pages. Nanonets bills each workflow step, with AI extraction at 0.30 credits per page. Open source libraries have no license fee. Our plans count files, not pages: $9 a month for 50, $24 for 500 with batch upload, and $49 for unlimited files plus the API.

Extracting data from PDFs you own or have received in the course of business, such as invoices, statements and reports, is routine and is what PDF scrapers are built for. Downloading PDFs from someone else's website is governed by that site's terms and by copyright, so check the terms before you automate collection. This is general information, not legal advice.

Scrape last month's folder and compare the totals

Take the PDFs you scraped by hand last month, run them through the tool at the top of this page, and tie the row counts and totals back to the originals. If they match, the batch upload and the API run on the same engine. Plans and yearly billing are on the pricing page.