I follow a handful of fields closely — AI, biotech, longevity research, space, startups. Keeping up with them properly turned into a second job, and I was still behind.
The problem was never a shortage of sources. It was the opposite.
Three things break a reading list
Volume. Subscribe to the feeds that actually matter in a field and you get hundreds of items a day. Most of them are noise, but you only learn which ones after you've read them.
Duplication. One real event produces dozens of retellings. A study gets published, and then you read the press release, the news write-up, the aggregator's summary of the write-up, and three blog posts reacting to the summary. Four minutes in, you realize you already knew this.
Language. A lot of what I want to know first shows up somewhere I don't read. Clinical results, regulatory decisions, and startup news tend to live in national press long before anyone writes about them in English.
Every reader tool I tried made the first problem worse and ignored the other two. More sources didn't make me better informed. It made me further behind.
What RealFeedApp does
It's a digest, not a reader. The point is not to show you everything — it's to show you what's worth your ten minutes.
- Collects from feeds, science preprints, and community sources across the topics you care about.
- Removes duplicates by meaning, not by URL. Forty retellings of one study collapse into one card, so a story you've already seen doesn't come back wearing a different headline.
- Ranks by significance, so a funding round doesn't outrank a clinical result just because it was published more recently.
- Writes a short AI summary for each item, in your language, so you can decide whether to open it.
- Always links to the original. Every card is a door to the source, not a replacement for it.
That last one matters to me more than it probably sounds. RealFeedApp doesn't republish anyone's article and doesn't try to be where you finish reading. It's the layer that decides what deserves your click. The source keeps the reader, the traffic, and the credit.
One technical note, since this is DEV: the summaries are generated by a model I run on my own server, not through a third-party API. Nothing you read passes through someone else's inference endpoint on the way to you.
Who it's for
People who need to track a fast-moving field and don't have an hour a day for it — researchers, founders, investors, and the kind of engineer who reads arXiv on purpose. I built the first version for exactly one user, which was me, and only then made it something other people could open.
Status
It's in beta and free. No signup, no install — open the link and the feed is there. Topics are filterable; summaries are available in more than one language. An Android app is in progress.
You can try it here: realfeedapp.com
What I'd like to know
If you open it, I'd genuinely like to hear:
- What's missing? Which source in your field should be in there and isn't?
- Do you trust the summaries? Specifically — did any summary make you skip an article you'd have wanted to read, or open one that wasn't worth it?
- Is the deduplication too aggressive? If two items look like the same story to the system but aren't to you, that's the bug I most want to hear about.
Feedback in the comments is welcome, and it goes straight into what I build next.
Top comments (0)