This is pulling the content of the RSS feeds of several news sites into the context window of an LLM and then asking it to summarize news items into articles and fill in the blanks?
I'm asking because that is what it looks like, but AI / LLMs are not specifically mentioned in this blog post, they just say news are 'generated' under the 'News in your language' heading, which seems to imply that is what they are doing.
I'm a little skeptical towards the approach, when you ask an LLM to point to 'sources' for the information it outputs, as far as I know there is no guarantee that those are correct – and it does seem like sometimes they just use pure LLM output, as no sources are cited, or it's quoted as 'common knowledge'.
https://github.com/kagisearch/kite-public/issues/97
There's also a line at the bottom of the about page at https://kite.kagi.com/about that says "Summaries may contain errors. Please verify important information."
1. It seems to omit key facts from most stories.
2. No economic value is returned to the sources doing the original reporting. This is not okay.
3. If your summary device makes a mistake, and it will, you are absolutely on the hook for libel.
There seem to be some misunderstandings about what news is and what’s makes it well-executed. It’s not the average, it’s the deepest and most accurate reporting. If anyone from the Kagi team wants to discuss, I’m a paying member and I know this field really, really well.
A lot of times when I ask for a source, I get broken links. I'm not sure if the links existed at one point, or if the LLM is just hallucinating where it thinks a link should exist. CDN libraries, for example. Or sources to specific laws.
It actually seems more like an aggregator (like ground.news) to me. And pretty much every single sentence cites the original article(s).
There are nice summaries within an article. I think what they mean is that they generate a meta-article after combining the rest of them. There's nothing novel here.
But the presentation of the meta-article and publishing once a day feel like great features.
So if this automates the process of fetching the top news from a static list of news sites and summarizing the content in a specific structure, there's not much that can go wrong there. There's a very small chance that the LLM would hallucinate when asked to summarize a relatively short amount of text.
When you go to Google News, the way they group together stories is AI (pre-LLM technology). Kagi is merely taking it one step further.
I agree with your concern. I see this as a convenient grouping, and if any interests me I can skip reading the LLM summary and just click on the sources they provide (making it similar to Google News).
A) redacted the news in a format that is read friendly
B) set up a page with prioritized news
Because _that’s what a newspaper is_.
What extra value is gotten from a AI rewrite? At best is a borderline noop, at worst a lossy transformation (?)
Services listing sources, like Kagi news, perplexity and others don't do that. They start with known links and run LLMs on that content. They don't ask LLMs to come up with links based on the question.
> Privacy by design: Your reading habits belong to you. We don’t track, profile, or monetize your attention. You remain the customer and not the product.
But the person running the LLM surely does.
That’s not news. That’s news-adjacent random slop.
Far more interesting is how they aggregate the data. I thought many sources moved behind paywalls already.
Imagine if Google news use LLM to show summaries to the users without explicitly saying it's AI on the UI.
Ironically, one of the first LLM-induced mistakes experienced by average people was a news summary: https://www.bbc.com/news/articles/cge93de21n0o.amp
I might not agree with all decisions Kagi makes, but this is gold. Endless scrolling is a big indicator that you're a consumer not a customer.
Someone recently highlighted the shift from social networks to social media in a way I'd never thought about:
>> The shift from social networks to social media was subtle, and insidious. Social networks, systems where you talk to your friends, are okay (probably). Social media, where you consume content selected by an algorithm, is not. (immibis https://news.ycombinator.com/item?id=45403867)
Specifically, in the same way that insufficient supply of mortgage securities (there's a finite number of mortgages) led to synthetic CDOs [0] in order to artificially boost supply of something there was a market for.
Social media and 24/7 news (read: shoving content from strangers into your eyeballs) are the synthetic CDOs of content, with about the same underlying utility.
There is in fact a finite amount of individually useful content per unit of time.
[0] If you want the Michael Lewis-esque primer on CDOs https://m.youtube.com/watch?v=A25EUhZGBws
Now I just read the news on a Sunday (unless I'm doing something much more exciting). For the remainder of the week I don't read the news at all. It's the way my grandad used to read the news when he was a farmer.
I've found it to be a convenient format. It let's you stay informed, while it gives enough of a gap for news stories to develop and mature (unless they happen the day before). There's less speculation and rumours, and more established details, and it has reduced my day-to-day stress.
Annoyingly I still hear news from people around me, but I try to tune it out in the moment. I can't believe I used to consume news differently and it baffles me why I hear of people reading/watching/listening to the news 10+ times per day, including first thing when they awaken and last thing before they sleep. Our brains were not designed for this sort of thing.
I would agree that a single daily news update is useful (and healthy), but this must also be reflected in the choice of topics and the type of reporting.
Summaries are no substitute for real articles, even if they're generated by hand (and these apparently are not). Summaries are bound to strip the information of context, important details and analysis. There's also no accountability for the contents.
Sure, there are links to the actual articles, but let's not kid ourselves that most people are going to read them. Why would they need a summarizing service otherwise? Especially if there are 20 sources of varying quality.
There are no "lifehacks" to getting informed. I'll be harsh: this service strikes me as informationally illiterate person's idea of what getting informed is like.
Should all politicians' remarks be reproduced verbatim with absolutely no commentary, no fact-checking and no context? Should an article about an airplane crossing the Pacific include "some experts believe that this is impossible because Earth is flat?"
Excessive bias in media is definitely a problem, but I don't think that completely unbiased media can exist while still being useful. In my expierence, people looking for it either haven't thought about it deeply enough, or they just want information that doesn't make their side look bad.
I agree, but how do you envision that happening? Journalism died a long time ago, arguably around the birth of the 24-hour news cycle, and it was further buried by social media. A niche tech company can only provide a better way to consume what's out there, not solve such large societal problems.
> There are no "lifehacks" to getting informed.
I don't think their intent is to change how people are informed. What this aims to do is replace endless doomscrolling on sites that are incentivized to rob us of our attention and data, with spending a few minutes a day to get a sense of general events around the world. If something piques your interest, you can visit the linked sources, or research the event elsewhere. But as a way of getting a quick general overview of what's going on, I think it's great.
FWIW, I agree with you.
I used to be a news junkie. I've always thought of writing the lessons I learned, but one of them was "If you're a casual news reader, you are likely more misinformed than the one who doesn't read any news." One either should abstain or go all in.
I guess I'd amend it to put people who only glance at headlines to be even more misinformed. It was not at all unusual for me to read articles where the content just plain disagreed with the headline!
(I was very skeptical about Kagi Assistant but now i am a happy Kagi Ultimate subscriber).
I like that Kagi charges for their service, so their motive is to provide services for that cost, and not with ads on top of it.
What I actually want is a curated set of things that are useful to me personally given my situation. The most important things about my situation to give me useful news are things like: net worth, income, citizenship, family situation, where I live, what industries I work in, current investments, travel destinations, regulatory and political risks associated with any of those things, etc.
Because those are the things that dictate how the parts of the world I can't control are going to affect me (especially if I don't react). I don't want to hear about random things that aren't going to affect me when I'm looking at the news. Sometimes I want to learn new random/useless things for fun, but that's a leisure activity. It's totally separate from the "news", which is a thing that adults consume as a chore to better plan their lives.
The fundamental problem is that myself and others are not going to willing give out the personal information required to curate useful news feeds, so the news will always be filled with noise. Maybe local AI can help with that.
I know the announcement page talks about not scraping, but to me personally the value i see in this product is that i don't have to go to those ad ridden, poorly organized and often terrible pages of the authors. Which then seems really unfair to the actual content providers.
I'd like to see this type of service cost $3-5/m ontop of my normal Kagi sub to compensate the authors of the articles i read. A Streaming Music model for news, ish.
This proposed value is quite small, but my assumption is only a very small amount of money would reach them from my ad views anyway so a $10/m addition feels extreme to me.
Could you guys maybe print it on paper and send it to my physical mailbox, so I can do this ritual with breakfast? :-)
Guten: A Tiny Newspaper Printer - https://news.ycombinator.com/item?id=42599599 - January 2025 (106 comments)
Getting my daily news from a dot matrix printer - https://news.ycombinator.com/item?id=41742210 - October 2024 (253 comments)
This is awful. It's cutting out any money going to the news agencies that go out there and write news. If they didn't exist, Kagi wouldn't work.
Why would Kagi stop working if news didn't exist? Kagi is a search engine first and foremost, Kagi News is not a money making product of theirs. Kagi would still be making money with their search engine.
Also, this should entice news writers to write better news. The main reason people use products such as this is that they are sick and tired of going to news sites only to have to power through filler material to get the 10% that actually matters...
News Minimalist [1] and Boring Report [2]. Both aggregate news and (IMO) most importantly provide links from multiple outlets for the same stories. Really made me notice the clickbait and allows me to be more selective in choosing reputable sources.
Both use AI, with the former ranking news based on importance, while the latter summarizes articles. (That doesn't feel useful for supporting journalism as a whole so I typically click through and read the articles unless I don't like the outlet reporting)
> Both use AI, with the former ranking news based on importance
I like this! If I'm in a rush, I check for very high priority stories. Usually there are 3 or even none. Done!
On days I want to sit back and read, it provides nice sources.
That said, I do think the service could be improved. Often the summary is a very short blurb that forces me to go to one of the original sites for the content, and hopefully land on one that is not obnoxious to use, which kind of defeats the purpose. The event timeline sounds interesting, but when it essentially shows 2 or 3 events that are obvious from the context, it's not so useful in practice. I always skip the "Quick questions" section, since it reads like an elementary school report, and the questions are really basic. How about letting me ask the questions I want?
Also:
> We don’t scrape content from websites. Instead, we use publicly available RSS feeds that publishers choose to provide.
I think this is a mistake. Most publishers are hostile to RSS and often don't offer it. Scraping is, unfortunately, a requirement if you want to consume public content on your own terms, which is the entire point of this service. Besides, scraping is how all search engines generate their index, so as long as the bot is well behaved and doesn't hammer the site, follows robots.txt or perhaps even bends the rules a bit, it should be fine. I would rather Kagi wasn't so respectful of publishers' wishes, if that would allow them to offer a better service. I understand if they want to avoid getting in trouble with publishers, but the alternative would be better for their users.
Nice release nonetheless!
The UK section seems to have a heavy bias towards news from Scotland.
It looks too simplistic for me to actually use.
If you wanted to fix the news you'd begin by critically curating mainstream news and throwing 80% of it in the trash, then you'd add 80% of material and critical analysis back to the 20% that had none of that.
I do wish I could have better control of what languages I'm getting. Right now the option is to either translate everything or nothing. I'd prefer news in their original, untranslated form if it's one of the 4 languages I speak, otherwise translate them to English.
I added the category "Israel" and everything was in Hebrew, so I had to set my language to English, but now news in my native Swedish are translated to English and I have to kind of translate it back in my head as I read them.
It's not the end of the world, but it seems like fairly low-hanging fruit!
"Mark as read" checks all the checkmarks, but since they're still there after a reload, I don't see the point.
I think keeping them on the page instead of automatically hiding them makes more sense for a product that's trying to update their news feed once per day. You feel more in control, as if it's not a stream of never-ending stories, but rather a fixed amount of stories that you can realistically power through. Seeing all items checked sort-of supports this philosophy.
A save feature to keep track of interesting articles would be nice.
Having more news (or more filtered for quality) would also be nice. Right now at 12 the lists seem to be mostly taken up by trendy low-quality news that will be irrelevant and less news that doesn't make waves but will probably have more impact in the long run. Actually this might just be a lack of the number of places being scraped. Not an actual example from the site but consider how much an article of someone saying the latest comet is actually alien technology trends (but is completely irrelevant) vs a scientific paper reporting on the measurements of the atmospheric composition of a bunch of exoplanets.
If you live in big city beware that your newspaper probably is lacking your neighborhood coverage which is what you need.
Yet, there is Hacker Newsletter (https://hackernewsletter.com/, which I like and use), there are others pointed by GPT5 that I don't Mailbrew and Digest. Kagi looks like the true former.
What I do want is personalization - not by picking interest, but actual personality, prompt, tastes, good enough that it puts something other, rather than only narrowing and narrowing my view. Yet high quality, rather than clickbaits and other "fluff". Otherwise, following a few Reddits would do the job (with some API to send emails).
What I would like even more is something that actually turns my social media into daily emails.
If you missed a day of news, whatever was really important will re-surface in today's news (major world incident)
Otherwise, perhaps what was missed is noise!
This example includes a Reddit post as a source:
https://kite.kagi.com/s/hjgy55
But that post is actually a link to reuters.com
There is also a list of "citations" which are referenced from the generated text, and "sources" which are not referenced anywhere. It's not clear if they used reddit or reuters to generate any of the text.
I also see lots of citations to "common knowledge"... which is um, weird.
For example:
> National Guard activation: Guard forces can serve under state control (Title 32) or be federalized (Title 10), which determines who directs missions and the scope of authority [*].
Is this common knowledge?
About "common knowledge" sources - we validate all content for accuracy. When the LLM needs to add context that's missing from sources (e.g. historical background), we mark these as "common knowledge" since this generated content can't be validated against the original sources. You're right that your example isn't common knowledge at all, we'll work on adding actual sources for these claims too.
Thanks for trying it out!
I do however like the fact that Kagi only pushes _once_ a day. Drinking from the firehose is physically and mentally exhausting. Even daily feels like too much these days other than a quick check to make sure the world didn't implode or the Rapture happened while I was busy trying to get CC to behave.
Gives me a good high-level view of the news. I'm a Kagi customer and I definitely don't want anything they do with the news.
If you haven't seen it there's also an amazing feature that you can go back and see the homepage as it was from any point in time in the last 20 years
Can you expand on why?
It's just plain text web 1.0 page that uses some ranking algo to figure out the top stores of a given day across categories, and shows that headline and under it similar headlines across different news sources.
It used to pull in RSS from the sources so you could also read the articles in plaintext, but that broke a bit ago and the dev hasn't fixed it.
Regardless, I still find it a great site to quickly get up to speed on top stories of the day!
But also I really like (and pay for!) Kagi so happily support their own effort here.
That's despite the appropriate HTTP header:
Accept-Language: en-US,en;q=0.5
When you share a Russian story with a non-russian speaker, they will still be able to read the story in their own set Content Language in the Settings. We're working on improving the UX of language, sharing a story, and more.
However, I set my feed up on the web app, seeing that it should sync on "all my devices".
Next, I installed the Android app, and mybe I missed something, but I don't see any way to connect to my Kagi account.
So much for syncing...
Some UX friction i noticed: To get back to the homepage from an article, i have to click on the article headline. While this is elegant and you likely get used to it, once you know it, it's not exactly intuitive.
Miniflux (https://miniflux.app/) in Docker, fetching 75 RSS feeds I've collected over the years
~200 lines in a Jupyter notebook:
- Fetch entries from Miniflux API (last 24-48 hours)
- Convert to CSV, feed to LLM. GPT-5 identifies trending stories across sources
- Each article gets web-fetched and summarized via Gemini-2.5-flash
- Results render via IPython.display
Ten minutes per day, fully informed.Apart from that, it's really nice! Good job, kagi team!
It feels to me like the bigger problem is more about assembling time series of "news" not "news today".
Like if you wanted "show me all stories about crime X from the BBC since 1980" or whatever but then you want to do this across many sources.
This is the missing piece for most new analytics. I think there are legal blockers to getting this done and why I mention decentralization.
I had a little trouble imagining myself using this in particular but I'm a big fan of the search engine.
Some things that could change that:
- Deep fact checking. Community Notes on twitter do a better job at this than any other system I've seen. The reason it doesn't really work in practice is that the stream of misinformation and confusion is orders of magnitude larger than the Community Notes community. A news app should not have that scalability issue.
- Follow up. If I read something that later turns out to be false I need to be notified of that. This unfortunately requires that the app track what I have read.
- Context. If you have a news article about a stabbing, it sounds like stabbings are up. The context that they are going up or down statistically is extremely relevant. The lack of context can turn a tiny truth into a bigger lie.
- Deep confusion analysis. Figuring out where people are confused statistically and focusing on trying to manage that misinformation gap is not something that is dealt with at all. I would like to become LESS confused by information sources not more.
News is broken because journalism is no longer a viable career path. No amount of RSS aggregators will fix that.
I hoping this can fill a gap for me currently. I want something that will give me broad awareness of big news I should probably know about, that’s not a 24 hour firehose of news.
I like the once-per-day update and the relatively short list of stories. The jury is still out on how sticky it will be, in terms of being my go-to place for a daily update.
I'm currently working on a major overhaul to provide more holistic context around news by better surfacing less-discussed events.
I like that it only provides the list once a day (I do think that's a clever feature), but the inability to influence bias seems like a mistake, especially since the sources already seem to follow a bias.
You have a great search service. Please focus on that. Build that into an actual Google-beater. Provide the features your customers actually want. Spend your time, money, and energy making that the greatest search service possible.
Don't waste this opportunity. Please.
Mozilla fell into this trap as its business model was fundamentally broken (majority revenue coming from biggest compatitor). Our business model is healthy and the more apps we have in the ecosystem the stronger the ecosystem gets.
This sounds like it's going to be a massive headache. Activists with nothing to do all day will be all over this, for their chance to try to have influence over what other people read.
I've been really enjoying Semafor's emails too, but their 2x a day is tough for me to keep up with. I'll try to get a habit of looking at Kagi News to stay informed.
I think it is human curated, but I'm not positive about that.
I tended towards Axios but lately it's gotten a bit paywalled and less informative. Can't wait to incorporate Kagi News into my daily workflow.
An attitude of "Hook me up to the novelty juice, this is old weak sauce", is a sign of internet / news addiction.
The news feature feels a bit underwhelming and underdeveloped though, especially with the LLM/AI approach.
These guys are doing great work and this news product is exactly what I want... Once a day hit. What is happening in the world? As far as pmf goes they hit the mark for an old fart like me.
(Edit) Now I see. You have to scroll through the story and click "Close story" to get back. It's "mobile first".
So far, i quite enjoy having a summary with bullet points.
For example, here's the summary of this discussion: https://extraakt.com/extraakts/kagi-s-daily-news-ritual-spar...
* for now
Bunch of discussion here 3 months ago? https://news.ycombinator.com/item?id=44518473
It was in beta then.
Sort of like a loss leader, eg the Costco hot dog :-)
You have defined the desirable news as "pure, essential information". What's that again? How do you know what's pure and essential info for any user? The traditional news media had started there, with that pure news, and ended up here where they are today.
Ultimately, you will realize that your content need to grab attention enough so that people consume your feed. People's attention goes to where things look weird, exciting, sensational, emotional, trivia, gossip etc. You can't do away with all that and just dish out the pure and essential info. It didn't work. People tried it.
- Parquet de Paris ouvre 24 enquêtes pour menaces
- Update: famille et experte ADN au procès Jubillar
- Intersyndicale appelle à la grève du 2 octobre
This won't be used by French speakers as is.
Let me open the app once a month and see a summary of what has happened over it.
RSS works great and there are a million ways to consume them. There are also a myriad New aggregator offerings, most with some sort of LLM thrown in on top.
Did we really need this? Was there nothing better Kagi could dedicate its resources to?
To be fair this is exactly how Kagi Search happened too - many people didn't see a point in a paid search engine in 2018 too, but I and my family needed one and it happened.
But the Sports section is bad. The game finished 10 hours ago and it's still showing a match preview.
can't imagine it would go over well in the court system.
- Allow me to have a single feed (as opposed to one tab/feed per category). Also, to prevent that feed from becoming too long, allow me to set a maximum number of news items or maximum number of minutes I'd like to spend. Prioritize/leave out news items accordingly. In other words: While I might be interested in sports, I'm not interested in reading or scrolling through as many news items about sports as about, say, world politics.
- "Highlights" and "perspectives" below the article text read like useless AI slop that merely reiterates the text, and artificially prolong an otherwise neatly concise page.
- Allow me to intersect categories and/or choose a regional "focus". Non-regional categories like "sports", "business", "technology" currently seem to aggregate news from across the world. However, I might be particularly¹ interested in a regional subset of e.g. business or sports news.
¹) I.e. not exclusively so. I'm still interested in world news but only when it comes to major events (in the sports case, say, world cups and championships).
While I understand different people find value in different things, dismissing Kagi generally as "too expensive" is ignorant IMO.
I think it also depends what you use it for. I use both their search and their AI models for software development and it saves me precious time when looking for information - in a way it pays for itself.
I think upvoting/downvoting is a crucial aspect to news/information/knowledge. But we've been doing it with just numbers all along. Why not experiment with weights or more complex voting methods? Ex: my reputation is divided in categories - I'm more an expert in history then politics hence my vote towards historical subjects have more weights. Feels like that's the next big step for news. Instead of just another centralized aggregator?
No offense to the cool system and website though
I feel this is what Apple News should've been. Instead it's just god-awful ad-filled mess of news articles. And the only reason I have it is because of Apple One. But it is a clearly neglected product.
I also pay for ground news but it hasn't met my expectations, mostly because there's a lot of redundancy with wire stories. Like it'll show 50 sources but they're all just regurgitating the same AP or Reuters article. So it skews the "bias"
Are these articulated in a manner which gives stakeholders (investors, users, and staff) assurances and standing?
...
What are competitors and collaborators in this space? Semafor seems to have a similar product, what are the differentiators and/or collaboration opportunities?
...
Netflix was subscription only, till it was "pay to get rid of ads". Then there is the whole business of profiling customer interest, etc.
We have product labeling for food, why not web services?
For me, this is only useful as a curated list of news feeds (and subreddits I guess), but nothing more.
[1]: https://github.com/kagisearch/kite-public/issues/97#issuecom...
[2]: https://kite.kagi.com/about
[3]: https://github.com/kagisearch/kite-public/blob/main/core_fee...
[1] https://embit.ca
Lately I've been working toward less app time and more boredom https://youtu.be/orQKfIXMiA8?si=ZyvxO0SFjoGGHbdK ¯\_(ツ)_/¯ works wonders
Parasitic by definition.
And embracing the news from nowhere perspective.
So both a parasite and boring at the same time.
I wish more tech folks who want to "fix the news" would learn from Gabe Rivera's Techmeme, Memeorandum, and Mediagazer.
He's done aggregation right for 20 years
It seems like all their recent releases are just following into the AI hype.
Trump, Congress deadlock as shutdown deadline nears
Taliban cuts internet nationwide, flights grounded in Afghanistan
Indonesia school collapse leaves 38 missing, 77 hurt
YouTube settles Trump suspension lawsuit for $24.5m
German court jails AfD aide for China spying
US deports 120 Iranians after deal
Russian drone strike kills family of four
Is this really what I need to know in the world? Am I saying “informed”? This is not helping the anxiety from reading news described in the article. This is not good for people.