Microsoft Publisher is gone.
If you had it through Microsoft 365, access ended on or shortly after October 1, 2026, as the removal update rolled out. Support for the perpetual version ended October 13, 2026, when Office LTSC 2021 and Consumer Office 2021 reached end of support. Perpetual installs keep running, but they get no more security updates.
Your files are not deleted. That part matters, so I want to be clear about it: your .pub files are still on your computer. What died is the reader — the program that could open them.
I found that out the way a lot of people are finding it out now, looking at folders of old newsletters, flyers, and brochures that nothing on the machine will open properly any more.
The official escape hatch has a catch
Microsoft's own guidance was to convert files to PDF inside Publisher before the deadline (File → Save as → PDF), and Microsoft also offered a bulk path: a downloadable script, Convert-PubFileToPDF.ps1.
The catch is in the script's first step: it needs a working, licensed copy of Publisher installed.
So for Microsoft 365 users, the bulk tool dies with the access it was meant to outlive. It only serves perpetual owners while their installs last. If you already lost access, it cannot help you.
Microsoft says the same thing on its support page, in its own way: it points users to third-party conversion tools, with a caution that such services vary in quality and capabilities, and that Microsoft provides no support for them. In writing, the vendor is handing its stranded users to the third-party market.
The real risk isn't conversion. It's silent failure.
I've spent time actually measuring what happens when you convert .pub files with the open tools available, and the problem isn't that conversion fails loudly. It's that it can appear to succeed while dropping things.
In the measurements behind this post, one converter run exited successfully and produced no PDF at all for a truncated file. Another file was converted and reported fine, but the extracted text came out reordered and garbled — "Sp ot Bu the nnn y & E gg ns" — while looking converted. Graphics in other files exploded into hundreds of small image tiles inside the PDF: 501 images in a single one-page flyer in one case, 70 in another.
All five test files were reported as successes by a naive conversion. Not all five were actually complete.
That's the risk I care about. A church newsletter archive or a school flyer collection that "converts" with pages, text, or images quietly missing is worse than a file that fails openly, because you might delete the originals believing you're safe.
The free option, stated honestly
LibreOffice opens .pub files, and The Document Foundation has promoted that support as a rescue path. If you have one file, a working computer, and you're comfortable installing and running desktop software yourself, try that first. It's free.
Use it with your eyes open: layout can shift, especially on graphics- and WordArt-heavy files, and you still have to check the result page by page to be sure nothing was dropped. Free and real — but it's you doing the checking.
What I'm testing
I build small pay-per-call tools, and I'm testing whether there's a real need for this as a service:
Send a .pub file, get back a PDF.
$0.10 per file, one at a time.
$0.05 per file in batches.
Every conversion completeness-checked before it's handed back: pages present, text present, images present, checked against the source file.
That's a structural completeness check. It is honestly not a pixel-perfect promise. I can't promise the PDF looks pixel-identical to what Publisher would have rendered — Publisher itself no longer exists as the ground truth to compare against. What I can check, and report, is whether the pages, the words, and the images survived the conversion, and flag the files that need a closer look.
The ask
I'm not building this on a guess. There's a short waitlist page here:
https://exchangerebel.com/publisher-rescue/
It asks two questions:
How many .pub files do you hold?
Would you pay per file to convert them?
There's also an optional email field if you want to hear the result, and an optional note if there's something I should know about your files.
If enough people sign up in the next two weeks, I'll build it. If not, I won't — either way I'll post the result here. That's the deal with a test like this: the answer is the answer, and I'll share it either way.
— DataRebel
Top comments (0)