All posts
Guides

My Reading Workflow: RSS to Obsidian, With Nothing Lost in Between

N
Nick VogelAugust 1, 2026 · 4 min read
Subscribe

Here's my process in one paragraph. I read my feeds in Riverbed a few times a day. When something's worth keeping, I save it with a sentence about why. That save becomes a markdown file in my second brain, an Obsidian vault, where an AI assistant sorts it and enriches it for research, future writing, or quick reference. Reading in, markdown out, nothing lost in between.

That's the workflow. The rest of this post is how each piece works day to day.

When I actually read

I open my reader a few times a day. If I need a brain break from coding, or I'm just hanging out on the couch, that's a great time to open Riverbed and swipe through a couple of articles. I use it on my phone as a progressive web app and on my computer.

I don't spend energy deciding what gets read now versus later. A few trusted sources I'll read immediately. If something's interesting but needs more investigation, I save it. If the headline doesn't do anything for me, I skip it. That's my entire triage system, and it keeps me informed without getting overwhelmed.

The save

When something's worth keeping, I use Riverbed's save-with-why. The why-line is a sentence or two, just long enough for me to remember why it was worth coming back to. I always write one because the app doesn't let a save through without a reason. That constraint sounds arbitrary, but it's why the system works. A save with a why is a note, and a save without one is a bookmark I know I'll never open again.

From there, Riverbed writes a markdown file of the full content, important information about the article included, to my vault through an outbox on a schedule. I don't copy-and-paste anything, I don't clip anything, I don't share-sheet anything.

Where it ends up: an inbox I never touch

Everything goes to an inbox folder, and I don't touch it after that. My AI assistant sorts through the inbox using my why-lines, asks questions when it isn't sure what goes where, and files each save into a research library. I don't worry about organizing it "right," because there's a set of written instructions for how sorting works, and when something feels off, I think it through and update the instructions.

I may not have a video ready to make or a blog post to write right then, but I know I will. When that day comes, my second brain ties the related topics together, and I come in and add the original thought. Reading I did months ago keeps paying off.

What I used before (the graveyard)

Before I had a second brain, saved things were just lost to the ether, or they'd end up in Apple Notes, buried in a Safari read-later list, lost somewhere in an unstructured vault, or forgotten entirely. Once the second brain existed, I ran the Obsidian web clipper and the iOS share sheet for a while. That worked, but it needed me to do the moving every time. I wanted that streamlined, so I built it.

The weak point of Riverbed

The process is tied to Obsidian, and that's a real limitation, not a footnote. The outbox writes markdown, and today the client that picks it up is the Obsidian plugin, so this workflow asks you to be an Obsidian person first. My own vault is synced, and I still know that sync is the single thread the archive hangs on. If I lost it, the reading would keep working and the keeping would stop, which is exactly backwards from what I built this for.

So I'd like more destinations over time, Notion among them, and a hosted option in Riverbed itself for people who don't keep a vault at all. That's the direction this is heading. I'm building what I need first and widening it from there, and I'd rather tell you the shortfalls plainly than pretend they aren't there.

If you're starting with 500 bookmarks

First, be honest about which of the 500 you'll actually read. You saved them once, so maybe there's something useful in there, and maybe not. Either way: feed the list to your AI assistant, let it interview you to figure out what's still important to you, delete the rest, and enrich what remains. Then start this loop going forward, reading by feed, saving with a reason, and letting the vault fill up with things you can actually find again.

New posts land in the feed first. Subscribe.
Join the waitlist