How to Save a LinkedIn Carousel or Document Post for Your AI Knowledge Base

How to Save a LinkedIn Carousel or Document Post for Your AI Knowledge Base

If you've ever tried to save a LinkedIn "carousel" — one of those swipeable slide-deck posts full of frameworks, hot takes, or case studies — for later research, you've probably hit a wall. There's no obvious "save" button on the post itself, screenshotting ten slides one at a time is tedious, and even if you manage to grab the images, you're left with pictures of text instead of something an AI tool can actually read. This guide covers both halves of that problem: how to get the file off LinkedIn in the first place, and what to do with it afterward so tools like NotebookLM, ChatGPT, or Claude can genuinely use it — with PostToSource handling the parts LinkedIn doesn't.

What a LinkedIn Document Post (Carousel) Actually Is

LinkedIn calls these "document posts" internally, even though almost everyone refers to them as carousels. A creator uploads a PDF, PowerPoint, or Word file, and LinkedIn renders each page as a swipeable slide directly in the feed. They've become one of the highest-engagement formats on the platform — dense frameworks, "here's what I learned" breakdowns, and slide-style case studies all get posted this way, which makes them a genuinely useful source for competitive research, industry tracking, or building a personal swipe file. The problem is that a carousel isn't a link you can paste anywhere, and it isn't plain text either — it's a rendered document sitting inside LinkedIn's feed UI.

The Good News: LinkedIn Already Lets You Download the Original File

Unlike a normal LinkedIn text post, document posts actually have a built-in export path — most people just never find it. Per LinkedIn's own help documentation (checked 2026-09-07), the steps are:

1. Open the document post and view it in full-screen mode.
2. Click the Download icon in the top-right corner.

That gives you the original file the poster uploaded — PDF, PPT, PPTX, DOC, or DOCX, per LinkedIn's supported formats. LinkedIn also caps what creators can upload in the first place: 100MB or less, and 300 pages or less per document, per the same help page. So if you're ever wondering why a huge slide deck won't post, that's the ceiling.

This native download only works from LinkedIn's desktop site, and it only applies to actual document posts — not a plain image post or a single-image "carousel-style" graphic, which LinkedIn treats as an image, not a document, and doesn't offer the same download button for.

The Catch: A Downloaded Carousel Often Isn't AI-Ready

Here's where the real work starts. A lot of LinkedIn carousels are built in Canva or a similar design tool, exported as a PDF of flattened slide images rather than a document with a real text layer. Open one in a PDF viewer and try to select a sentence — on plenty of carousels, nothing highlights, because there's no extractable text underneath the visuals. Paste that file into NotebookLM or a ChatGPT project and you'll either get nothing useful back or an answer built on shaky OCR guesses instead of the actual words on the slide.

The other gap: the downloaded file is just the slides. It doesn't include the post's caption, the poster's name, the date, or the top comments — all context that matters if you're using this as a research source later and want to remember why you saved it or who said it.

This is the exact gap PostToSource is built to close. Paste the LinkedIn post URL in and PostToSource pulls the caption, author, and post context into a clean, properly-formatted file alongside the visual content — text-first, not an image dump — so it's actually searchable and citable once it's a source in your AI tool of choice.

Step-by-Step: From LinkedIn Carousel to AI Knowledge Base Source

1. Find the carousel and open it full-screen on LinkedIn's desktop site.
2. Use the native Download icon to grab the original file if you want the exact slide visuals preserved.
3. Check if the text is selectable — try highlighting a line in your PDF viewer. If it selects cleanly, you likely have a real text layer and can upload it as-is.
4. If the text won't select (a flattened image export), convert the post's LinkedIn URL with PostToSource instead — you'll get the caption and post content as clean, AI-readable text, which covers the substance even when the slide visuals themselves aren't extractable.
5. Add it to your knowledge base alongside your other converted sources — X threads, newsletter issues, other LinkedIn posts — so a single NotebookLM notebook or ChatGPT project can reason across all of them together. If you're stacking a lot of sources, it's worth knowing where each AI tool's upload limits sit before you hit a wall mid-research.

Real Workflows This Fits

• Competitive tracking: Marketers saving competitors' carousel frameworks weekly to spot messaging shifts, then asking an AI tool to summarize the pattern across a quarter's worth of posts.
• Personal swipe files: Creators collecting high-performing carousel formats from their niche, then querying NotebookLM for "what structure do the highest-engagement posts share?"
• Research roundups: Analysts pulling slide-based data or case studies from LinkedIn into the same knowledge base as newsletters and X threads on the same topic, rather than treating each source type separately.

For the rest of your LinkedIn research stack — regular text posts, articles, and newsletters — see how to convert standard LinkedIn posts to PDF and the broader LinkedIn posts to AI knowledge base workflow. If you're building out a full research library across platforms, PostToSource's guide to the best NotebookLM sources and the guide to importing social posts into NotebookLM are good next stops.

Frequently Asked Questions

Can I download a carousel someone else posted, or only my own?

The native Download button works on any public document post you can view in full-screen mode, not just your own posts. LinkedIn doesn't restrict downloading to the original poster.

Why won't NotebookLM or ChatGPT read the text in my downloaded carousel PDF?

Most likely the carousel was exported as flattened slide images (common with Canva-style designs) rather than a document with real text underneath. AI tools can still process image-only PDFs in some cases, but results are far less reliable than a proper text layer — test by trying to highlight text in the PDF first.

What's the file size or page limit for LinkedIn document posts?

Per LinkedIn's own help documentation (checked 2026-09-07), uploaded documents must be 100MB or less and 300 pages or less, in PDF, PPT, PPTX, DOC, or DOCX format. That's the ceiling for what a creator can post — not a limit on what you can save afterward.

What if the post is just a single image, not a multi-slide document?

If it's a single-image post rather than a true document post, LinkedIn won't show the Download icon, since it's classified as an image, not a document. In that case, convert the post's URL directly with PostToSource to capture the caption and any visible text as a clean source instead.

Whether the slides come with a real text layer or not, the goal is the same: get the substance of what was said into a format your AI tools can actually search, cite, and reason over — not just a folder of screenshots.