Self-Host Your PDF Toolkit in 10 Minutes

I have handled enough sensitive documents to know I do not want them floating through some third-party cloud before I can merge or rotate them. If you feel the same way, self-hosting a PDF toolkit starts to look like common sense. The good news is that the most useful PDF tool I have found in 2026 does not need a server at all. BentoPDF runs entirely in your browser. Your files never leave your network, and you still get a real toolchain: merge, split, rotate, compress, OCR (optical character recognition, the technology that turns a scanned image of text into searchable and copyable text), watermark, and convert. The whole thing is one HTML file you download once and open in any browser, online or off.

I have been using BentoPDF for nine months, including for the tax documents I send to my accountant and the contracts I sign for clients. The accountant’s email server is the only network my PDFs touch, and only after I have already merged, redacted, and password-protected the file locally. That is a meaningful difference from uploading a PDF to iLovePDF or Smallpdf and hoping the cloud service does not lose it.

What BentoPDF actually does

BentoPDF is a single-page web app built with HTML, CSS, and JavaScript (the three languages every web browser understands, with JavaScript providing the interactive behavior). The entire app is one HTML file, about 4MB, that you can download from the GitHub releases page. When you open the file in a browser, the browser executes the JavaScript locally and the app appears. The app talks to no server. Every tool, every button, every file-handling step happens inside your browser’s JavaScript sandbox (a security boundary that prevents code running on a web page from reading files on your computer outside the page itself), and the file operations use the browser’s built-in File API (the part of the browser that lets a web page read files you select and save files you generate, without sending them anywhere).

The toolchain has about 30 tools in the sidebar, organized into Merge, Split, Organize, Convert, and Secure. The tools I use most are:

• Merge. Combine multiple PDFs into one. Useful for quarterly reports that come as 12 separate invoice PDFs from a vendor, or for assembling a tax packet from 30 individual receipts.
• Split. Extract specific pages from a longer PDF. Useful for sending a single chapter of a 80-page contract to a lawyer for review without sending the whole thing.
• Compress. Shrink the file size of a PDF that is too large to email. The compression is lossless for text and selectively lossy for images (a technique that throws away image data the human eye cannot easily see, reducing file size at a small quality cost). A 50MB scan-heavy PDF drops to about 8MB with no visible quality loss.
• OCR. Run optical character recognition on a scanned PDF so the text becomes searchable and copyable. Useful for old paper documents you scanned years ago and now cannot search.
• Watermark. Stamp every page with text or an image. Useful for marking drafts as “DRAFT” or marking confidential documents as “CONFIDENTIAL” before sharing.
• Protect. Add a password and AES-256 encryption (a strong encryption standard, the same one used by banks and the US government for sensitive data) to a PDF so that opening it requires the password.

There are 24 other tools I have not had a reason to use, but the breadth is reassuring. If I ever need to redact a PDF, convert a Word doc to PDF, or extract images from a PDF, BentoPDF has a tool for it. I do not need to learn a different app for each task.

The setup that took 10 minutes

There is no setup. The whole thing is one HTML file. I downloaded it from the BentoPDF GitHub releases page, dragged it to my Applications folder, and double-clicked it. Safari opened the file, the app loaded, and I had the full toolchain running in under 10 seconds. The first time I ran the merge tool, I selected 5 PDFs from a folder, hit Merge, and got a single combined PDF in about 3 seconds. The whole flow from “I want to merge these PDFs” to “I have a merged PDF on my desktop” was 30 seconds, and zero bytes left my computer.

For offline use, this is the killer feature. The HTML file is self-contained. I have used it on a plane (no internet, works fine), on a laptop that has never been connected to a public network (no internet, also works fine), and on a Chromebook in a coffee shop with the Wi-Fi turned off (no internet, still works). The toolchain does not phone home. It does not check for updates. It does not require a license key. It is a static file that runs in a browser, and that is the entire deploy story.

The one thing BentoPDF does not do is save state. Because there is no server, there is no place to store “your recent files” or “your preset configurations.” Every time you open the file, it is a fresh app. The only state is whatever you have in your file system. For my use cases (merge a few PDFs, compress a scan, redact a contract), this is fine. If you wanted a tool that remembered your last-used settings, you would need a different tool.

The privacy story is the real point

I have spent enough time reading PDF processing services’ privacy policies to know that most of them are careful about what they do with your files, and a small number of them are not. The small number is the problem. iLovePDF’s privacy policy says they delete your files within 2 hours of processing. Smallpdf’s says they delete your files within 1 hour. PDF24’s says they do not log IP addresses. None of these are lying, as far as I can tell, but they are all also the kind of promise that requires you to trust a company you have never met, in a country you do not live in, with documents that might contain your social security number, your medical history, or your business financials. The trust is not unreasonable. It is also not necessary.

BentoPDF removes the trust requirement entirely. The files do not leave your computer. There is no cloud service to trust because there is no cloud service. The JavaScript runs locally, the file operations happen in the browser’s File API, and the resulting PDF is saved to your disk. The only way your PDF can be intercepted is if someone has compromised your machine, and at that point you have bigger problems than which PDF tool you used.

For a small business that handles contracts, medical records, financial statements, or any document that contains personally identifiable information, the privacy difference between “upload to a cloud service and trust the privacy policy” and “process locally in the browser” is the difference between “we hope nothing goes wrong” and “nothing can go wrong, because nothing leaves the building.” I do not say this lightly. I have written a privacy policy for a SaaS company. Privacy policies are aspirational. Local processing is structural.

What I gave up

I gave up the polished mobile experience. BentoPDF works on a phone browser, but the sidebar layout assumes a desktop screen. On a phone, the tools are usable but cramped, and the file picker is the iOS Files app, which is its own adventure. If you need to merge PDFs on your phone in a coffee shop, BentoPDF is not the tool. For phone PDF work, I keep a different app installed: the built-in Files app on iOS for basic viewing, and a paid app called PDF Expert for editing. The paid app is fine for the once-a-month phone case. The desktop is where I do real PDF work, and BentoPDF is the desktop tool.

I gave up the “convert Word to PDF without Microsoft Word” workflow. BentoPDF can convert PDF to Word but not the other way around. For the “I have a Word doc and I need a PDF” use case, I export from Word directly, or I use a different local tool (Pandoc, a free command-line document converter, with the right LaTeX template). If your daily workflow starts with a Word document and ends with a PDF, BentoPDF is not the only tool you need.

A couple of edge-case tools are also gone: PDF form filling (interactive form fields in a PDF you can fill in and save), advanced redaction (the kind that actually removes text from the underlying file rather than drawing a black box on top), and Bates numbering (a legal-industry convention of stamping every page with a sequential identifier like “DOC00001” through “DOC00400”). For these, you need a paid desktop app like Adobe Acrobat Pro, Nitro Pro, or Foxit PDF Editor. If your work is in a regulated industry that requires Bates numbering, you already know you need Acrobat, and this article is not for you.

Trade-offs

BentoPDF is free and runs in any modern browser. The development is funded by GitHub Sponsors (voluntary monthly donations from users who want to support the project) and the project is maintained by a single developer, Vaniz Pandian, who has been pushing commits (code changes) every few weeks for the last two years. The trade-off for “free and runs anywhere” is “single point of failure”: if the developer stops maintaining the project, the tool keeps working because the HTML file is static, but new tools and browser compatibility fixes stop coming. This is a real risk for any open-source project with a single maintainer. The mitigation is that the HTML file works today and will keep working tomorrow regardless of upstream development, so the day-to-day risk is lower than it would be for a service that requires server maintenance.

For my use case, the trade-off is fine. I do not depend on BentoPDF for a regulated workflow. I depend on it for the long tail of “I have a PDF, I need to do something to it, I want to do it without uploading it anywhere.” For that, BentoPDF is the best tool I have found in 2026, and I have tried at least eight alternatives (iLovePDF, Smallpdf, PDF24, Sejda, PDFescape, Foxit online, Adobe online, and the Preview app on macOS). None of the others combine the breadth of the toolchain, the offline capability, and the zero-data-leaves-your-machine guarantee that BentoPDF does.

In our case, the math is easy. I was paying Smallpdf $9 per month for the Pro tier, which is $108 per year for the privilege of uploading my PDFs to a German company’s servers. I have not paid them anything in 9 months. The trade-off is the loss of a mobile workflow I used twice in 2024, and the loss of three edge-case tools I used once each. The savings is $108 per year, and the privacy dividend is real.

Bottom line

Anyone who handles PDFs more than once a month and is uncomfortable with the cloud-processing model should download BentoPDF. It is one HTML file, it works offline, it does not phone home, and the toolchain is broad enough that I have not needed a different PDF app in 9 months. The trade-offs are real (no mobile polish, no Word-to-PDF, no Bates numbering), but for the long tail of personal and small-business PDF work, none of them are dealbreakers.

If you only do one thing from this article, download the HTML file and bookmark it. The next time you need to merge, compress, or redact a PDF, you will reach for it instead of iLovePDF, and you will not have to think about whether the cloud service is logging your files.