June 1, 2026 · Updated September 27, 2026 · 8 min read · PDFPrivacy

The Best Free PDF Tools in 2026 That Work Entirely in Your Browser

Hitendra Patel

Founder, keptlocal · Senior Technical Lead, Healthcare IT

Most "free" PDF tools have a catch: they upload your files to a server, process them there, and send the result back. The privacy policy says files are deleted after an hour. You have no way to verify that. For a personal photo, this trade-off is fine. For a contract, a medical record, or an HR document — it is a real risk you should know about before hitting upload.

Browser-based PDF tools are different. They run the processing code directly in your browser tab using JavaScript and native browser APIs. Your file never travels across the network. The computation happens on your own CPU. You can verify this yourself: open DevTools → Network tab, then run the tool. The tab stays quiet.

This article covers what browser-based PDF tools can genuinely do in 2026, which tools do it well, and where the limits are.

What browser-based PDF tools can do today

The assumption that serious PDF work requires a server was true in 2015. It is no longer true. Modern browsers expose a full set of APIs that make client-side PDF processing practical:

pdf-lib — a pure JavaScript library that reads and writes PDF documents. Handles merging, splitting, rotating, reordering, watermarking, adding page numbers, deleting pages, and PDF compression — including, as of a recent update, recompressing the embedded photos that usually make a PDF large, not just stripping metadata.
pdf.js — Mozilla's PDF rendering engine, the same one inside Firefox. Renders PDF pages to canvas at any resolution. Powers PDF-to-JPG and PDF-to-PNG conversion. Also extracts text content for PDF-to-text tools.
Canvas API + FileReader — handles all image processing: resize, crop, rotate, flip, grayscale, format conversion, Base64 encoding, and (via the browser's own JPEG encoder) recompressing photos at a lower quality.
WebAssembly — lets compiled C/C++ code run in the browser at near-native speed. Some browser-based tools use it for heavier decoding work (HEIC photo decoding is a common case); plenty of the operations above run in plain JavaScript instead.

The result: the large majority of everyday PDF work can now run entirely client-side, including the "make this smaller" job people usually assume needs a server. The parts that genuinely still don't: PDF-to-Word conversion, OCR, and true multi-pass Ghostscript-level compression for extreme cases. I built keptlocal around that split: local-first for everything that can be local, and honest about the boundary where it can't.

The best browser-based PDF tools in 2026

Tool	Processing	Free tier limits	Signup	Ads (free tier)
keptlocal	Your browser	None — no file size or usage cap	Not required	None
Smallpdf	Their servers	About 2 tasks/day on the free plan¹	Not required for basic use	—
iLovePDF	Their servers	Per-tool file caps (15–400 MB) and monthly quotas, e.g. 25 merges, 2 compressions²	Not required for basic use	Yes
PDF24	Their servers (a separate desktop app is local)	No official cap stated³; independent reviews report a roughly 100 MB practical ceiling⁴	Not required	Yes
Squoosh (images only)	Your browser	None stated⁵	Not required	None

Checked 2026-09-27. Free-tier limits and terms change; verify current terms on each tool's own site before relying on them. ¹ reported by G2's Smallpdf pricing page, not stated on Smallpdf's own site at the time of checking. ² per iLovePDF's own pricing page. ³ per PDF24's own site ("no artificial limits"). ⁴ per an independent PDF24 review. ⁵ per Squoosh's own site.

keptlocal.com — built privacy-first from the ground up

keptlocal.com is a purpose-built collection of 37 browser-only tools for PDF and image work. Every tool in the collection runs client-side — no tool in the PDF or image category uploads your file to a server.

On the PDF side: Merge, Split, Rotate, Reorder Pages, Watermark, Delete Pages, Add Page Numbers, PDF to JPG, PDF to PNG, PDF to Text, Compress PDF, Unlock PDF, JPG to PDF, PDF Info Viewer, and more. On the image side: Compress Image, Resize Image, Crop Image, Convert Image (HEIC to JPG, PNG to WebP, and more), Rotate Image, Flip Image, Grayscale Image, Image to Base64, and more. See the full tool list for everything currently live.

The site is free, requires no account, and has no watermark on any output. The design is straightforward — no upsell banners, no "upgrade to Pro" interruptions, no daily limits.

Smallpdf.com — strong tool depth, aggressive upsell

Smallpdf has a broad feature set, including PDF-to-Word and OCR. The catch: independent pricing trackers report the free tier limits you to around two tasks per day, and many operations (including PDF-to-Word) require a paid account. Files are processed server-side. The privacy policy states files are held for a limited time, but that is a policy claim, not a verifiable technical guarantee.

Good choice if you need PDF-to-Word or OCR and are comfortable with server-side processing. Poor choice for frequent use on the free tier — the daily task limit is intentionally restrictive.

iLovePDF.com — large-scale, upload-based

iLovePDF is the market leader by traffic (over 250M monthly visits as of early 2026, per Similarweb — a figure that will drift over time). It covers almost everything. Files are uploaded to their servers. The free tier's limits vary by tool rather than following one blanket cap: per iLovePDF's own pricing page, file-size ceilings range from 15 MB to 400 MB depending on the operation, plus monthly quotas on some tools (for example 25 merges or 2 compressions). Ads appear throughout the free interface. The paid Premium tier starts at roughly €4–7/month.

Good choice if you need heavy-duty operations (true PDF compression, PDF → Office conversion) and you are comfortable with server-side processing. The free tier's per-tool quotas can be frustrating for repeat use on a single operation.

PDF24.org — comprehensive, but cluttered

PDF24 is free with no tier system, which is genuinely rare — PDF24 states there are no artificial limits, though independent reviews report a practical ceiling of around 100 MB per file for the web tools. Like Smallpdf and iLovePDF, the web version processes server-side and is ad-supported; unlike them, PDF24 also offers a free Windows desktop app that processes files locally. The web interface is cluttered and the mobile experience is poor, but the tool coverage is deep. It is the best choice if you want free-forever, server-side processing without a tier system.

Squoosh (images only) — Google's browser-based image compressor

Squoosh is the gold standard for browser-based image compression. It runs entirely in your browser (Google-built, open-source), supports AVIF and WebP encoding, and lets you compare the compressed result side-by-side with the original before downloading. It does one thing — compress images — but it does it exceptionally well.

How to verify a tool is genuinely browser-only

Do not take the marketing copy at face value. The verification takes 30 seconds:

Open the tool page in your browser.
Open DevTools: press F12 (Windows/Linux) or Cmd+Option+I (Mac).
Switch to the Network tab.
Drop a file into the tool and run the operation.
Watch the Network tab. A genuine browser-only tool shows no POST request with your file data. You may see requests for CDN scripts or fonts (those are fine) — but the tool's operation should not send your file bytes anywhere.

This test is definitive. A tool cannot upload your file without that upload appearing in the Network tab.

When browser-side processing is not enough

There are operations that genuinely require server-side processing for quality results:

Extreme PDF compression, or unusual image types: browser-based tools (including keptlocal's Compress PDF, which now recompresses embedded JPEG and RGB/grayscale images alongside stripping metadata) handle the common case well. What they still hand off to a server or desktop tool: CMYK print-workflow images, transparent images with soft masks, and squeezing a 50 MB scan down to the absolute minimum via multi-pass Ghostscript-level optimization.
PDF to Word (.docx) conversion: this requires understanding the PDF's layout geometry and reconstructing a Word-compatible document structure — a hard problem that no current client-side library handles well. Tools that claim to do this in the browser typically produce poor results on anything other than simple text PDFs.
OCR (optical character recognition): recognizing text in a scanned PDF image requires a heavyweight model. Tesseract.js can run in the browser but is slow and less accurate than server-side alternatives for multi-page documents.
Sophisticated redaction: properly removing sensitive text from a PDF requires understanding the PDF content stream — not just drawing a black rectangle over it (a naive approach that leaves the underlying text selectable). This is a solved problem on the server with tools like pikepdf; it is not yet solved client-side.

For these operations, a server-side tool is the appropriate choice. The honest framing is: use browser-only tools for the 80% of PDF work that benefits from privacy guarantees, and use server-side tools deliberately for the 20% that genuinely requires them.

Performance: does browser-based processing feel slow?

Modern laptops and phones are fast. For typical document work — a 10-page contract, a 20-image photo PDF, a set of receipts — browser-based processing is essentially instant. You will not notice a difference compared to a tool that uploads, processes on a fast server, and sends the result back.

The performance difference appears at extremes: a 500-page scanned PDF with high-resolution images, or a batch of 100 photos. For those cases, a server with more RAM and a GPU-accelerated encoder will outperform a laptop browser. For the work most people actually do most of the time, local processing is fast enough to be unremarkable.

Which tool to use

For everyday PDF and image work — merging, splitting, rotating, converting, compressing, cropping — browser-based tools are the right choice in 2026. They are as fast as server-based tools for typical document sizes, free to use, and structurally safer for sensitive documents because there is no third-party server involved.

Keep a server-side tool bookmarked for the operations that genuinely still need it: PDF-to-Word conversion, OCR, and the most extreme compression cases. For everything else — including compressing PDFs full of ordinary photos and scans — the browser is fast enough that you will not notice the difference, and the file stays on your device.

keptlocal.com covers the browser-only operations. For PDF-to-Word conversion or OCR, PDF24 (free, server-side) or Smallpdf (paid, server-side) are the best complements.