I stopped uploading sensitive PDFs to Adobe after finding a self-hosted tool that does everything Acrobat charges for

The PDF files that are worth protecting tend to be the boring ones, like a scan of your ID that an application needs, or a bank statement where an account number has to be blacked out. The free Acrobat Reader app is fine for reading and signing things on your own machine, but tools like redaction are locked behind the Pro subscription at $20 a month.

Anything past those basics goes through Adobe's free online tools if you don't want to pay, and those work by uploading your file to Adobe first. That leaves a slightly annoying choice between paying every month to keep the heavier work on your own PC or handing over the exact document you were trying to make private.

Stirling PDF is the free open-source option that usually comes up when people complain about this. And its 3.0 release last month added a bunch of new automation features too, so it seemed like a decent excuse to finally get it running and try it out.

The everyday PDF work Acrobat puts behind a paywall

The free version of Acrobat can't touch some of these

Stirling PDF actually crams a lot into an app that has quite a minimalistic look. The docs name 55+ tools, and they're sorted in the sidebar under headers like signing and document security. Many of these tools do overlap with what Acrobat already does for free, like filling in forms for example, but signing has a nice extra since you can scan a QR code and draw your signature on your phone.

The Editor is where it starts to cover paid ground because editing existing text in Acrobat needs at least a Standard plan at $15 monthly (on an annual plan). Stirling's editor was stuck in alpha for a long time, which meant it was still an unfinished, early-testing feature, but the 3.0 drop rebuilt it and now it's finally out of alpha. You click into any text on the PDF and just type. The one downside is that if the PDF's embedded font is missing a character you type, it swaps to Helvetica by default, which will probably look a little off.

Merging and compressing PDF files are technically free in Acrobat, but only through Adobe's online tools, so the file has to be uploaded first. Doing one locally via the desktop app requires a subscription. In Stirling, both are in the quick-access list and they're free. Moreover Redaction and Compare are Pro-only in Acrobat, whereas Stirling's Redact can find every instance of a word and black it out, or let you draw boxes over things manually, and Compare highlights the text differences between two versions of a PDF. There's also the Sanitize tool, which strips JavaScript and embedded files out - this is pretty handy for files you may have downloaded from a site you don't fully trust. Stirling also has OCR, which is another Pro feature in Acrobat.

What the 3.0 update actually added

Folders that sort themselves and an agent that reads along with you

Automation isn't new to Stirling. The Automate tool was already in V2, and it lets you save a chain of tools to run on any file. One of the suggested presets, for example, cleans out risky content and then runs OCR.

The 3.0 release went a lot bigger with something called the PDF Processor. It takes those same tools and runs them across huge batches of files from folders or FTP or S3 storage, then keeps listening for new ones. The full Processor dashboard only lives in the browser version, though.

In the desktop app you get a "process a folder" button instead. This lets you pick a folder and a preset, like Classification for example, and the app watches over that folder while it's open, sending new PDFs off for processing. Classification seems to be on in the background by default too, which is apparently how my library started tagging files as HR without me asking - my best guess is that the desktop onboarding did it.

Then there's the Stirling agent. You can ask it questions about any PDF you have open and it will pull directly from the file, and you can also tell it to run tools, like "merge these into one PDF". You can hook it up to a local model, to an API, or just use Stirling's own AI infrastructure - this uses a combination of Claude and VoyageAI models, and you get 1,000 credits every month which I think is plenty.

Keeping your PDFs on your machine

Two ways to cut out the cloud

There are two ways to keep your workflow local with Stirling. First up is local mode - there's no button for this, you just start using the app without signing in on first launch. Every tool then runs on your computer and nothing gets uploaded anywhere. A handful of the tools won't run locally, though, and the app will mark those with a small cloud icon. When you're signed in, those get sent to Stirling's servers for processing, while in Local Only they're blocked, which rules out OCR and converting Office files.

The other option is self-hosting Stirling with Docker. OCR comes back this way and you can also point the AI at your local models through Ollama or other OpenAI-compatible servers. As far as I can tell it's the only way to get the 3.0 automation without Stirling's cloud involved.

Turns out, I never needed Acrobat

Funny thing is that the setting I'll probably touch the most here is the logout button when I want to enter local mode for private documents, and I can still get AI by connecting it to my local Qwen. And whenever I do want those cloud-only features, I get to have them for free. I think there's still a case for Acrobat if your whole job lives in PDFs, but paying $20 a month to edit files I'd rather keep off the internet doesn't really add up.