The future of Linkstash

So hey, Linkstash isn’t dead.
But there isn’t much development happening right now.

I’ve been running my own instance of Linkstash and using it as my main bookmarking service for a while now. And having used it, I’ve discovered what I actually want out of it — and what I don’t.

A quick look at the origin

I made Linkstash because I needed a bookmark manager that fit me.

I’ve mentioned Pinboard and Linkding a few times as inspiration. They’re great, but there were just some small things I wished they did differently, so I made my own.

I never set out to compete with those tools. But I felt taht a little courtersy to the craft is in order, so I wanted to do it right. That meant writing docs, polishing the UI (UX really isn’t my strong suit, but I tried). Even if I’m the only user.

What I thought mattered (but didn’t)

After I finished the MVP features, I had a a list of things I thought would be important or interesting:

1. AI-based auto-tagging

This sounded cool — especially with all the AI hype right now.
But it turns out I never actually wanted it.

I always know why I’m saving a link. And being the kind of link hoarder that I am, I’ve already got my own tagging system.

I also didn’t like the idea of increasing the Docker footprint. Keeping Linkstash lean was always a goal. I hate when a simple selfhosted service has a huge Docker Compose file with 5-container.

In fact, I wanted to ship LinksStash with SQLite, but ended up going with MariaDB to support a few DB triggers that made some logic easier.

I still think some kind of rule-based auto-tagging might be worth exploring later — but for now, it’s just not a priority.

2. Semantic search

This wasn’t just AI hype — I genuinely thought it could help.

But in practice, tag filters and plain search have always been enough. I can almost always find what I need. A lightweight vector search might still be worth looking into, but again, it’s just not that important to me.

3. Tools for handling huge bookmark collections

A lot of work went into the bookmarks page, and I’m happy with where it landed.

The in-place filtering, sorting, and permalinks are exactly what I need. The bulk tag editing handles the rest.

Some vanity stats (like total bookmarks) might be fun, but I don’t need them.

4. Client libraries

The idea made sense: frontend should talk to the backend via a client library. Architecturally clean, opens up extending via different client.

But there’s no actual need for it right now.
I might revisit it later for fun, but it doesn’t help the user much at this point.

What actually matters (now that I’m using it)

1. Better CSS for archived pages

I thought the default archive CSS was decent. I even though it look quite good some of my test pages during development.

But as I used it more, I’m seeing that there are way too many sites that render poorly or are just unreadable in archive view.

And I’m not talking about weird edge cases — I mean regular articles and text-based content. Web pages are complex now, and it takes more work than I thought to make archived pages consistently readable.

This is going to be my first area of focus.

2. Better integration

I really want to:

• Stash a PDF and have it sent to my Paperless instance
• Stash a YouTube video and have it auto-sent to my Pinchflat instance

I feel like this kind of integration with other selfhosted tool would be very useful. Beyond those two, I don’t see many other integrations that I want to integrate with. These are the relevant ones that I run. But even just those would be a good quality-of-life improvement for me.

3. Mobile experience

Mobile layout is just isn’t there yet.

I did some work to make it responsive, but there are still screens with horizontal scrollbars and layout issues.

More importantly, there’s no good way to “share to Linkstash” on mobile. I want to support the native share integration so you can stash links directly from any app.

In summary

Linkstash is still alive, just not active.

I still use it every day. When I run into friction, I fix the friction.
That’s how it grows now.

This was always a personal project — in the real, old-school sense of the word.
I built it for myself. I still use it. I still improve it.

It’s not a startup. It’s not a business.
But it’s not abandoned either.

And if you’re one of the few who tried it, starred it, or just peeked at the README — thank you.
I’m genuinely grateful.

Thanks for reading.
— Farhan