TL;DR

There are three places a file-to-PDF conversion can happen: on a server (upload sites), in a desktop app (Acrobat, Preview, Word), or in your browser (a Chrome extension). Most people default to the first because it's the top search result, and it's usually the worst of the three.

Convert: Anything to PDF does it in the browser: pick files, get a PDF, nothing uploaded, no account, no watermark, no file size cap imposed by a free tier. It handles images (JPG, PNG, WebP, SVG, GIF, BMP), text and markup (TXT, HTML, Markdown, JSON, XML), CSV rendered as a proper table, and the active web page — and it merges several of those into one PDF.

It does not do OCR, and it does not parse .docx or .xlsx natively (Excel goes via a CSV export). If you need either of those, you need a different tool, and this post will say so rather than pretending otherwise.

The useful frame isn't "which tool is best." It's where do you want the conversion to happen, because that single choice determines whether your file touches a third party at all.


The three places conversion can happen

On a server (Smallpdf, iLovePDF, Adobe's online tools, PDFCrowd)

You upload a file. Their infrastructure converts it. You download the result.

Good at: format coverage. A server can run LibreOffice, Ghostscript, OCR engines, and font libraries. That's how these services genuinely convert .docx and run OCR on scans — real capabilities we don't have.

Bad at: everything about your file leaving your machine. It sits on their storage subject to their retention policy. It passes through their subprocessors. Free tiers commonly impose daily caps, file-size limits, and watermarks, and the paid tiers exist because conversion at scale costs money. And structurally, none of them can convert a page you're logged into — their server isn't you.

In a desktop app (Acrobat, Preview, Word, LibreOffice)

Local, capable, and usually the right answer for heavyweight document work.

Good at: the hard stuff. Real OCR, PDF/A output, redaction that actually removes content rather than covering it, form field editing, digital signatures, preflight.

Bad at: cost, weight, and friction. Acrobat is a subscription. Installing a full office suite to convert one Markdown file is absurd. And for anything that started life as a web page, a desktop app has no good story — you'd be printing from a browser anyway.

In the browser (a Chrome extension)

The conversion runs in the page process on your machine, using rendering the browser already does.

Good at: speed, privacy, and web-native formats. No upload, no account, no round trip. It's the natural home for anything that's already text, markup, an image, or a page — which, if you're honest about your week, is most of what you convert.

Bad at: anything requiring a binary format parser or a text-recognition engine. A browser doesn't ship a .docx reader or an OCR model, so an extension can't have one without downloading a large library or sending your file somewhere — and we're not doing either.

What Convert: Anything to PDF actually converts

The complete supported list, no padding:

CategoryFormatsTypical use
ImagesJPEG/JPG, PNG, WebP, SVG, GIF, BMPReceipts, screenshots, scanned photos, logos, diagrams
Text & markupTXT, HTML (local file), Markdown, JSON, XMLNotes, saved pages, docs, API payloads, config exports
TabularCSV → auto-formatted tableExports from a spreadsheet, database, or scraper
WebThe active browser tabAnything you're currently looking at

Two behaviours worth knowing:

SVG converts as vector. It isn't rasterized on the way in, so a logo or diagram stays crisp at any zoom in the output. That matters for brand assets and technical diagrams — more on that here.

CSV becomes a real table, with header styling and column alignment, not a wall of comma-separated text. If your table is wide, set landscape orientation and a larger paper size (A3 or Ledger) before converting; that's the standard fix for right-edge clipping, covered in fitting a wide table on one page.

Merging

Select multiple files and they become one PDF in the order you choose. This is the operation that most often drives people to an upload site, and it's entirely unnecessary to leave your machine for it. Mixed types work — three JPGs, a CSV, and a Markdown file merge into a single document.

What it can't do, stated plainly

We'd rather you find this out here than after installing.

No OCR. A photo of a page becomes an image inside a PDF. The text in it is not selectable, not searchable, not extractable. There's no optical character recognition anywhere in the pipeline. If you need searchable text out of a scan, you need Acrobat, macOS's built-in text recognition, or an OCR service.

No native .docx or .xlsx. These are ZIP containers full of XML with a large specification behind them. Parsing them faithfully in a browser extension means shipping a heavy library and still getting layout wrong. For Excel data, export to CSV and convert that — it's the documented path. For Word, use Word, LibreOffice, or Google Docs' own export.

No paywall or login bypass. For the web-page conversion, it captures what's rendered in your session. If you can see it, it converts. If you can't, nothing changes that.

No editing an existing PDF. This makes PDFs; it doesn't reflow, redact, or re-paginate one you already have.

Chromium only. Chrome, Edge, Brave, Arc, Opera, Vivaldi. Firefox and Safari implement different extension APIs.

Picking the right tool by task

TaskBest toolWhy
Photos or receipts → one PDFExtensionLocal, instant, merges natively
CSV → formatted table PDFExtensionRenders a real table; no upload
Markdown notes → PDFExtensionRenders formatting properly
JSON/XML → readable PDFExtensionPreserves structure and indentation
SVG logo → vector PDFExtensionStays vector
A web page → PDFExtension (Web to PDF)Only local tools see logged-in pages
Scanned document → searchable PDFDesktop app or OCR serviceRequires OCR; we have none
.docx → PDFWord / LibreOffice / Google DocsNeeds a real Word parser
Redacting sensitive contentAcrobat or similarTrue redaction removes data, not just covers it
PDF/A archival outputDesktop PDF softwareSpecific conformance requirements
Filling and signing a PDF formAcrobat, Preview, or an e-sign serviceForm field manipulation

The pattern: web-native and image formats belong in the browser; binary office formats and text recognition belong in a desktop app. Upload sites sit awkwardly in the middle — they can do the desktop-app jobs, but they charge your privacy for the browser-job convenience.

The logged-in-page problem nobody mentions

This deserves its own section because it's the clearest structural difference and it's routinely glossed over.

Ask an online converter to turn a page into a PDF and it fetches the URL from its servers. Its servers have no session on that site. So for an order confirmation, a bank statement view, an internal dashboard, a patient portal, a university records page — anything requiring a login — it gets a sign-in screen. Not a degraded result: the wrong page entirely.

A browser extension reads the page your authenticated session already rendered. That's not a clever trick, it's just where it runs. It's also why the extension can't do anything about content you don't have access to — it has exactly your access, no more.

The FAQ covers this for a long list of specific sites — Gmail, Notion, banking dashboards, Slack, and others — at the Convert: Web to PDF FAQ.

Verifying the "nothing is uploaded" claim yourself

Don't take our word for it. Ninety seconds:

  1. Open DevTools with F12 or Cmd+Option+I.
  2. Go to the Network tab.
  3. Clear the log.
  4. Run a conversion.
  5. Watch what appears.

You'll see the page's own requests. You won't see your file going anywhere, because it doesn't. Run the same test against any online converter and watch a multipart upload of your document leave the machine.

This test works on any extension, and it's the fastest way to check whether a "private, local" claim is real. We walked through the broader audit approach in auditing Chrome extension privacy disclosures.

What about Chrome's built-in Print to PDF?

Genuinely useful and already installed. Cmd+P → Save as PDF handles simple pages fine, and there's no reason to install anything if that's all you need.

Where it falls down:

  • It only converts pages, not files. It won't turn a folder of JPGs into a PDF.
  • It struggles with lazy-loaded images, capturing the pre-load state and leaving gaps.
  • It has no ad or clutter removal. Whatever's on the page is in the PDF, cookie banner included.
  • It repeats sticky headers on every page break.
  • It can't merge.

We compared them properly in Chrome Print to PDF vs. a dedicated extension. Short version: use the built-in for simple pages; use an extension when the page fights back or when you're converting files rather than pages.

Frequently asked questions

Is there a Chrome extension that converts files to PDF?

Yes. Convert: Anything to PDF converts images (JPG, PNG, WebP, SVG, GIF, BMP), text and markup (TXT, HTML, Markdown, JSON, XML), and CSV — plus the current web page — into PDFs entirely inside the browser, with no upload and no account.

Can I convert a file to PDF in Chrome without uploading it?

Yes, and that's the main reason to use an extension over a website. The conversion runs locally in your browser. You can confirm it with DevTools → Network: run a conversion and watch that no file leaves your machine.

Does it convert Word documents to PDF?

No. .docx isn't a supported input — parsing Word's format faithfully needs a full document engine that doesn't belong in a browser extension. Use Word, LibreOffice, or Google Docs' own PDF export instead.

Can it convert Excel spreadsheets?

Not directly. Export your sheet to CSV first, then convert the CSV — it renders as a properly formatted table with header styling. For wide tables, set landscape orientation and A3 or Ledger paper size to avoid clipping the right-hand columns.

Does it do OCR on scanned documents?

No. There's no optical character recognition in the extension. A scan converts into a PDF containing that image, and its text stays unsearchable. For searchable output from a scan you need Acrobat, macOS's built-in text recognition, or a dedicated OCR service.

Are there file size limits or watermarks?

No watermarks and no artificial size cap. The practical limit is your machine's memory, since everything runs locally. That's different from free tiers on upload sites, which impose daily conversion counts, per-file size limits, and watermarks by design.

Why can't online converters handle pages I'm logged into?

They fetch the URL from their own servers, which have no session on that site — so they receive the login page, not your content. Only a tool running inside your authenticated browser can see what you see.

Is it really free?

Yes. No trial, no premium tier, no account. Every feature is available on install.

Bottom line

Deciding where the conversion happens settles most of the question. Images, text, markup, CSV, and web pages are browser-native — converting them locally is faster, private, and free, and there's no reason to hand them to a server. Scans needing OCR and Office binaries need a real desktop application, and we'd rather point you there than fake it.

Upload sites make sense for the second category if you don't have the software and the file isn't sensitive. For the first category they're a bad trade you've been making out of habit.

Convert: Anything to PDF — 14 formats, merge support, no watermark, no account, nothing uploaded. Pair it with Convert: Web to PDF for pages, including the ones behind a login.

And for when the filing is done: CineMan AI shows IMDb and Rotten Tomatoes ratings right on Netflix, Prime Video, and Disney+.