My Confession About a “Capture Tool”

To be honest, at first I didn’t really want to admit that what I was building was a “capture tool.”

Those words sound too small. Small enough to disappear into the browser extension marketplace at any moment. Small enough that they don’t seem like the basis for a product worth talking seriously about. In this industry, we prefer bigger language. We talk about workflows, knowledge systems, AI-native products, the next interface. Compared to all that, “saving” sounds almost embarrassingly basic.

But the longer I worked on it, the more I felt that the real issue might be exactly there.

A lot of what actually shapes the quality of reading and thinking comes down to these tiny actions. You read something and want to keep it. In that one second, the way a system responds determines more than whether a feature feels smooth. It determines whether your train of thought survives the interruption.

My feeling about this didn’t come from theory. It came from being worn down by it, over and over again.

I’ve had that moment many times. I’d be reading a long essay about product thinking, come across a paragraph, and suddenly feel: yes, this is right, and it connects to a feature I’ve been thinking about. I wouldn’t want to just agree with it in passing. I’d want to keep it, maybe even add half a sentence of my own. But then the familiar sequence begins: leave the tab open and tell myself I’ll deal with it later; copy it into some notes app; throw it into bookmarks; open a web clipper, choose a destination, think of a title, decide whether it needs tags. By the time all of that is done and I return to the page, half of what was forming in my head is already gone.

There’s another version of this when I’m reading technical docs or papers. You’re not really “reading one thing” from top to bottom. You’re scanning, comparing, looking for the fragment that actually matters to the problem in front of you. What you want to save usually isn’t the whole page. It’s a definition, a short explanation, a diagram, an example. But a lot of tools aren’t really designed for that moment. They either just take the link away, or they send you wholesale into another system.

At some point I became very clear on this: we do not lack ways to save things. What we lack is a capture action that doesn’t interrupt us so much.

That sounds like a small sentence. But to me now, it feels more concrete than a lot of grand language about “knowledge management.”

Because the problem is no longer that there isn’t enough to read. The real condition is that there is far too much, more than anyone can fully process. Web pages, PDFs, forums, papers, docs, video transcripts, podcast summaries, and now layer after layer of AI-generated explanations, summaries, rewrites, and reductions. Information is no longer scarce. It’s ambient noise that keeps pouring in. Anyone can get a polished-looking summary in seconds, and anyone can rewrite a paragraph so it sounds understood.

That’s why I’ve stopped being impressed by the idea of “getting more information.”

The harder thing is whether you can hold onto that thin line of your own thought in the middle of all this noise. What made you pause, why you paused, which sentence made you want to keep thinking, these things are personal, and fragile. They are not just “knowledge points.” They’re closer to fingerprints left behind by thinking.

Part of why I’m building Vein is that I don’t want to lose those fingerprints.

I don’t really want to keep relying on those familiar temporary behaviors. Leave a pile of tabs open and tell yourself you’ll come back. Bookmark it for later. Drop it into something called inbox, as if the mere existence of an inbox guarantees that a more disciplined, clearer version of yourself will show up in the future and clean everything up. We all know how that usually ends. Your future self is not going to thank your present self for saving the 47th link. They’re just going to stare at a pile of contextless items and have no idea why any of them mattered.

That’s also why I’ve always felt that bookmarks are strangely honest. They expose a fact very clearly: saving a link is not the same thing as saving a judgment.

But the other side has its own problem. A lot of “more advanced” tools, in trying to solve the limits of bookmarks, swing too far in the other direction. They ask you to make too many decisions at the exact moment of capture. Where should it go? What should it be called? Which project does it belong to? Does it need tags? A template? These tools are not bad. Some of them are genuinely excellent, and I use them myself. But the most valuable moment in reading often can’t bear that much process.

To put it bluntly, a lot of tools make saving feel too much like admin work.

You were only trying to catch a thought, and the system handed you a form.

Vein grew out of that small but, to me, very real discomfort.

I’ve never thought of it as a “next-generation knowledge management tool.” That framing is too big, and it isn’t accurate. Vein is closer to a capture layer inside the browser. It doesn’t exist to finish all your organization for you. It doesn’t exist to build your entire knowledge system. It does one thing: when you come across something worth keeping on the page you’re already on, it tries not to let the act of saving pull you away from it.

I care a lot about the phrase “on the current page.”

Some people think that’s just an interface detail, whether it’s a sidebar or a new tab, a floating window or a pop-up. But the more I’ve thought about it, the more I feel it reflects a deeper difference in how a product understands reading.

There’s a class of tools that assumes real work only happens inside their own container. If you want to be serious, you’re supposed to enter their space, submit to their structure, and organize your material on their terms. That logic is common, and powerful. But reading and thinking are not always that tidy. A lot of what matters shows up before it has been organized. A good tool should help that moment survive first, instead of demanding that you immediately become the manager of your own knowledge base.

I don’t think every act of saving deserves a full classification ritual.

Most of the time, you only need three things: the original passage, the source, and one note from that moment. That’s enough. When you come back later, if you can tell where it came from, why you kept it, and what crossed your mind at the time, then the item is alive. Without that, even the most beautiful database can still be nothing more than a well-decorated warehouse.

So what Vein wants to do is not make structure heavier. It wants to make the minimum amount of structure that is still genuinely useful feel more natural.

And I think this matters even more in the age of AI.

Because more and more content will be automatically summarized, automatically categorized, automatically tagged. Models will get better and better at turning material into something that looks clean and usable. But I still believe the part that cannot really be outsourced is not information processing. It’s that moment of “why did this matter to me?” AI can help you extract what ten articles have in common. It cannot truly answer why this particular sentence, at this particular moment, hit you.

That matters to me a lot.

Not because of some romantic defense of the humanities, but because I increasingly feel that if a person’s knowledge system ends up containing only processed conclusions, and none of the traces left behind while those conclusions were forming, it gradually becomes flatter. Cleaner, maybe. More efficient, maybe. But less and less like your own mind.

Vein is not trying to stand against AI. I’m fully aware that AI will absolutely have a place in products like this. It can help extract, organize, connect, and resurface things extremely well. But I don’t want a product to put AI in front of the person from the very beginning, as if once the model is strong enough, reading is reduced to letting the system process everything for you. I still believe some parts of reading should belong to the person first, and only then be assisted by the machine.

That’s also why Vein is an independent product, and why it tries to stay local-first.

I know “local-first” can easily be treated as a posture now, something in the developer world that comes wrapped in an automatic halo of virtue. Or, in the age of vibe coding, a convenient badge for the irresponsibility of shipping half-finished products just because anyone can validate an idea at almost no cost. But for Vein, it isn’t a posture first. It’s an experience issue, and a trust issue. If a capture tool you use all the time has to ask the cloud for permission every single time, and you can’t even clearly say where your own saved material lives, it’s going to have a hard time becoming part of your thinking. You don’t hand your most immediate, most habitual action over to a black box you don’t fully trust. And as a middle layer, Vein only needs to do one thing well: capture. Cloud storage and syncing belong to downstream knowledge systems. Vein should stay as light as possible.

The state I want is simpler than that: it works offline, you know where the data is, and later, if you want to export, migrate, or reorganize it, you’re not trapped by some platform. Not because I need that to support some advanced ideological claim, but because I trust less and less the idea that it’s a good long-term bet to leave all your reading traces in the care of platform goodwill.

At the end of the day, Vein’s ambition isn’t about becoming big.

If it has any ambition at all, I want it to occupy a very early, very basic position. Not to become your only knowledge system, but to become the most natural action you reach for inside the browser. You see something useful, press a key, highlight, drag, jot half a sentence, and keep reading. Don’t let the action itself turn into a task switch.

I think that position is underrated today.

Everyone is building toward bigger systems, fuller closed loops, stronger automation. Very few people are really treating “how a person saves things while reading” as a problem worth refining on its own. But if that step isn’t done well, then even the most sophisticated downstream system is just receiving material that is mediocre, context-poor, and sometimes not even worth reopening.

In the end, what matters is often not how glamorous the downstream layer is, but whether the upstream layer caught the most important thing in the first place.

The reason I’m willing to spend time on Vein comes down to something simple: I’m exactly the kind of person who lives in the browser. I write code, look things up, read product writing, read comments, read technical docs, and run into genuinely illuminating things on all kinds of stray web pages. I know that feeling very well. It isn’t “I need to build a knowledge system.” It’s “there’s something here that I can’t let slide past me.”

That is the second I want this product to serve.

If someone asks me later what Vein is really trying to solve, I probably won’t use grand language. I’ll say this:

When you see something online that is worth keeping, saving it should not interrupt your ability to keep thinking.

That’s it.

But if that can really be done well, I think it gets closer to the core of reading than many flashier ideas do. Because reading has never been just about absorbing information. For anyone who reads seriously, it always involves filtering, pausing, associating, marking, doubting, and judging. If a tool can respect those processes instead of flattening them, interrupting them, or taking them over, then it has already done something rare.

This is what I hope Vein eventually becomes.

It doesn’t steal the spotlight. It doesn’t think for you. It doesn’t force you to organize too early. It doesn’t lock you inside some beautiful new system. It simply stays close to where reading happens and, when you need it, helps the things worth keeping stay kept.

If it can do that, I won’t feel like I built an ordinary “saving tool.”

I’ll feel like I added a missing layer of infrastructure to reading as it exists today, something that should have been there all along, but that almost nobody has seriously tried to build well.