I Built a News Aggregator in a Day

Every morning, before the noise of the day starts, I read the news. Not from one place. From many. I open one Nepali news site, then another, then another. By the time I have checked five or six of them, half an hour is gone and I am still not sure I saw everything.

For years I patched this with social media. A feed full of followed pages did the collecting for me. That patch died in September 2025, when Nepal switched off 26 platforms overnight and my morning routine broke along with everyone else’s. That was the week I stopped trusting rented shelves to hold my habits.

So I built my own shelf. On October 1, 2026, I launched Sisnoo, a small news aggregator for Nepal: one feed that pulls fresh headlines from every major publisher, plus a morning briefing called The Sisnoo Daily with five sections. Nepal, Politics, International, Finance, Technology.

It took a day. What it taught me is worth more than the product.

I was the first user

Every product I have built that worked started with a problem I had personally. Niguro, my search engine, started because Nepali search results did not understand Nepal. Sisnoo started because my own mornings were broken.

This matters more than any technology choice. When you are your own first user, you do not need user research to know what “good” looks like. You know when the feed feels right. You know which headline is stale. You can feel the product working before any metric tells you.

When I was shaping my book-writing workflow into a guide, I did not start with a table of contents. I started with the memory of people asking me, over and over, how I actually turn AI designs into real WordPress sites. The demand was already there in my inbox. Same rule here: I was the inbox.

If I had built Sisnoo for “the market,” I would still be in a planning document. I built it for my mornings. It shipped in a day.

Freshness is the whole product

Here is the thing nobody tells you about aggregators: the technology is easy and the freshness is everything.

A news aggregator that is two hours behind is not a slow aggregator. It is a museum. Nobody opens a news app to see what was important this morning. They open it to see what is important now.

So the real job of Sisnoo is not storing news. It is noticing new news, fast, and putting it in front of you before it goes stale. Everything in the build serves that one job. The feed refreshes constantly. The briefing goes out in the morning because morning is when fresh matters most.

This sounds obvious until you try to do it. Headlines age in minutes. A story that breaks at 7 in the morning is old by 10. If your pipeline checks sources once an hour, your readers are always reading history. I learned this the hard way years ago running Scriling, my news site: the days we published fast were the days traffic spiked, and the days we published late were the days nobody came.

Timestamps lie

The second thing I learned: you cannot trust the timestamp a feed hands you.

When you pull headlines from RSS feeds and news indexes, each item comes with a published time. Reasonable people assume that time is true. It is often not. A feed can show a fresh timestamp for an article that was actually published days ago, because what you are seeing is the time of the latest crawl, not the time of the writing.

I ran into this exact trap this week while testing a news automation on another project: the feed said a story was fresh, the publisher’s own page said it was from yesterday. I almost published something stale because I trusted the middleman.

The rule I now follow everywhere: the publisher’s own timestamp, printed on the article page itself, is the only ground truth. Everything else is a rumor about time. For Sisnoo this means the freshness logic reads the source of truth, not the echo. It is more work per headline. It is the difference between a feed you trust and a feed that embarrasses you.

The front door is often locked

The third lesson: getting the headlines in the first place is harder than it should be.

You would think the open web is open. In practice, the biggest news indexes block simple readers. Plain text fetches get refused. Redirect tokens that work in a browser fail from a plain HTTP client. You write what looks like a simple script, and it comes back with nothing.

The fix is unglamorous: read the publishers’ sites directly. No clever middleman, no shortcut through someone else’s index. Go to the source, read the page, take the headline and the timestamp, move on.

There is a lesson here that has nothing to do with code. Every shortcut in a pipeline is a dependency, and every dependency is someone else’s decision you cannot control. The week the platforms went dark in Nepal taught the whole country this. My aggregator taught it to me again in miniature. Build on the source, not on the middleman, whenever you can.

The same story, told ten times

Freshness is the hardest problem. The second hardest is duplication.

Ten publishers cover the same bus accident with ten different headlines. A naive aggregator shows you all ten, and your “one feed” becomes a wall of echoes. The reader came to you to save time. Showing them the same story ten times is theft of the time you promised to save.

Solving this well is real work: matching stories across publishers, picking the clearest headline, linking the rest. I will be honest that Sisnoo’s first version is simple here. It does the basics and improves weekly. But naming the problem on day one matters, because it keeps you from declaring victory too early.

This is where building in Nepal actually helps. The publisher set is small enough to reason about. Every major outlet, one list. You can hold the whole landscape in your head. Try doing that for a country ten times our size.

Boring is a feature

Sisnoo has no login. No comments. No personalization engine. No app to download. One feed, five sections in the morning briefing.

This was deliberate. I have built enough products to know that features are how builders avoid the real question, which is whether anyone wants the thing at all. The first version of Niguro was not a search engine. It was a news page that pulled headlines into one feed. One page, one job. People visited, so the bigger dream earned the right to be built. I wrote about that process in detail when I described how I decide what to build.

Sisnoo follows the same rule: the smallest version that teaches me something. Does anyone open an aggregator for Nepal every morning? The briefing is the test. If people read The Sisnoo Daily for a month, the product deserves more investment. If not, I learned something for the price of a day.

A day. That is the whole budget this product has spent so far. When the cost of learning is one day, you can afford to learn a lot.

The morning is the product

One decision I am proud of: the briefing goes out in the morning, and it has exactly five sections. Nepal, Politics, International, Finance, Technology.

News is not a stream. A stream never ends and never satisfies. News is a meal. You sit down, you eat, you are done, you go live your day. The endless scroll is a design choice made by companies that sell your attention. A morning briefing is a design choice made by someone who respects your attention.

This is the part of Sisnoo that is opinionated, and opinion is what makes a product worth using instead of merely usable. Anyone can show you headlines. The briefing says: here is what matters, in five buckets, now go have your day.

If you want to build one yourself

Here is my honest checklist, from a day of building and a week of fixing:

1. Be your own first user. If you would not open it every morning, nobody will.
2. Treat freshness as the product, not a feature. Check sources constantly. Trust only the publisher’s own timestamp.
3. Read publishers directly. Every middleman in your pipeline is a door that can lock.
4. Solve duplication before you polish design. Ten echoes of one story is a broken promise.
5. Ship the smallest version that teaches you something. One feed, one job. Features can wait.
6. Make it a habit, not a stream. A morning briefing respects attention. An endless feed harvests it.
7. Keep the publisher set small enough to hold in your head. Depth beats coverage.

Small tools compound

Sisnoo is a small thing. It will never be my biggest product. But small tools built for real mornings compound in ways big launches do not. Every lesson above now lives in Niguro’s news systems too. The freshness logic, the timestamp discipline, the dedup thinking. A day spent building Sisnoo made every future news product I build better.

That is the quiet math of building: you are never just building the thing in front of you. You are building the judgment you will use on the next ten things.

I opened Sisnoo this morning. Five sections, everything fresh, done in six minutes. Then I closed it and got to work. That is exactly what it is for.