How to Fill Web Forms from a PDF with AI

Here’s a familiar scene. Your resume is a PDF. The job application is a web form on Workday. Two windows side by side, and you retyping name, address, dates, and job titles field by field. Twenty minutes later you hit submit and hope you didn’t transpose a digit in your phone number.

PDFs are where copy-paste goes to die. The information you need is right there, laid out cleanly — but a web form can’t read a PDF, and your browser’s autofill sure can’t either. So you become the bridge, manually shuttling data from one to the other.

AI Form Fills now reads PDFs directly, so you can fill a form from a PDF without the retyping. Open the PDF, capture, open the form, fill. This post walks through how it works and where it actually helps.

Why PDFs break traditional autofill

Browser autofill and most form-filler extensions assume you have a static profile: name, email, phone, address, typed in once. That assumption falls apart with PDFs in three ways.

First, a PDF isn’t a form. There are no <input> fields to detect — just text on a page. Autofill has nothing to hook into. Second, every PDF is different. Your resume, a vendor’s W-9, an invoice from a supplier — none of them share a structure, so a static profile can’t hold them. Third, PDF layouts are unpredictable: two-column resumes, multi-page contracts, datasheets with dense tables. There’s no consistent shape to match against.

So the manual workflow sticks around. The fix isn’t a better profile; it’s a tool that can actually read the document.

How AI Form Fills reads a PDF

The extension uses pdf.js — the same engine Firefox uses to render PDFs — to read the document inside your browser. The PDF isn’t uploaded anywhere just to be parsed. Here’s what happens:

1. Detection. Open a PDF (a .pdf link, or Chrome’s built-in viewer) and the extension recognizes it from the URL.
2. Extraction. pdf.js walks the document page by page, reconstructing lines of text by their x/y position on the page. Two-column layouts come out in the right reading order, not jumbled.
3. Understanding. The AI reads the extracted text and turns it into structured data — names, addresses, dates, amounts, line items, whatever the document contains.
4. Filling. Open any web form, click Fill, and the AI matches the right piece of data to each field by meaning, not by label string-matching.

One wrinkle worth knowing: if the PDF is a local file:// that the extension can’t fetch directly, or it’s open in Chrome’s internal viewer, you’ll get a file-picker prompt to re-select the file. It’s a one-time step per session, and it exists because Chrome blocks extensions from reading those URLs directly. Slightly clunky, but it beats not being able to read local PDFs at all.

Filling a job application from a PDF resume

The most common scenario. You have a resume PDF and a Workday or Greenhouse application form.

Open the PDF. It can be a downloaded file, a PDF hosted online, or one open in Chrome’s viewer — all detected.

Capture. Click the extension icon. Add a prompt if you want to narrow things down, e.g. “extract work experience, education, and contact details — skip references.” The AI reads the PDF and holds the result as structured data you can reuse.

Open the form. Navigate to the application — ATS portal, company careers page, multi-step registration, whatever.

Fill. Click Fill. The AI reads each field’s label, placeholder, and surrounding context, then maps the right value in. For tricky fields, a prompt like “if there’s no clear match, leave it blank — don’t guess” keeps it honest.

Review and submit. A quick scan, then submit. The part that used to eat your afternoon takes a couple of minutes.

Where this actually pays off

Job applications. This is the obvious one. Capture your resume once and reuse it across every application in a session — different companies, different form layouts, no retyping. If you’re applying to dozens of roles, this alone justifies the install.

Vendor and supplier onboarding. Vendor portals constantly ask for company details, tax IDs, and contact info that already live in PDFs — W-9s, certificates of insurance, company profile sheets. Capture the PDF, fill the portal form.

Invoices into accounting systems. An invoice PDF lands in your inbox; you need its line items in your accounting tool or procurement portal. Capture the invoice, fill the entry form. Amounts, dates, PO numbers, vendor names — all transferred without typing them back in.

Government and insurance portals. Plenty of these still publish templates as PDFs while the actual submission happens online. Extract from the PDF, fill the web form.

Multi-page web forms work too — click Fill on each page and the AI keeps applying the right data to each section from the same captured source.

A note on scanned PDFs

pdf.js extracts the text layer embedded in the PDF. That covers anything exported from Word, Google Docs, Excel, or a PDF editor — basically most PDFs you’ll encounter in a business setting. Pure-image scans with no text layer (old faxes, photographed documents) won’t have anything to extract, and you’d need OCR first. If you capture a PDF and get back almost nothing, that’s almost certainly what happened.

PDF autofill vs. the old way

Getting started

1. Install the extension from the Chrome Web Store.
2. Open any PDF in your browser.
3. Capture what you need — one click, optional prompt.
4. Open the form, click Fill, review, submit.

The free plan covers 20 fills a month, which is plenty to put it through its paces on real applications. Pro removes the cap if you’re doing this at volume.

If PDFs have been the bottleneck in your form-filling workflow, this is the piece that was missing.