COMPARED
Best news aggregators that link to primary sources
Every aggregator promises the news. The question is where the headline sends you.
An aggregator that links out is a routing layer: it reads the sources, ranks them, and hands you off. An aggregator that links inward is a destination: it reads the sources and keeps you. Both are legitimate products and neither is hiding it, but the difference decides whether you can check a claim in one tap and whether the reporting still reaches you if the aggregator shuts down. The six below are compared on that axis first and on everything else second.
| Aggregator | Where a headline sends you | How items get picked | Cost |
|---|---|---|---|
| The Front Wire | The originating outlet or agency — and for seismic, weather and filings, the agency record itself | Ranked automatically from the feeds it watches; the same order for every reader | Free to read; $3/mo or $24/yr membership |
| Techmeme | The original report, with the outlet named beside every headline | Human editors make the final call and write the headlines, on top of automated crawling | No reader subscription; advertisers, data sales and news-filtering services |
| Ground News | One of the outlets merged into that story, picked by you from the list | Articles about the same event merged across outlets, presented with bias labels | A Subscribe option is offered; no price named on the About page |
| Readless | The original article — but after an AI summary of it, and only for sources you connected yourself | Nothing is picked from the open web; a model summarises and de-duplicates the newsletters, Substack posts and RSS feeds you added | $4.90/mo Pro, $9/mo Max; seven-day Pro trial, card required |
| Google News | Google’s own News surfaces first; Top stories in Search also links on to the News tab | Computer algorithms, with some sections personalised from your Google and YouTube activity | No fee named in Google’s own help pages |
| The raw public feeds | Nowhere — the feed is the primary source | Nothing is picked; you get everything the agency publishes | Free (USGS, NWS, SEC EDGAR) |
What “links to the primary source” actually means
There are two different things people mean by primary source, and an aggregator can do one without the other. The first is the publisher: you tap a headline and land on the outlet that reported it, on their site, under their masthead. The second is the record itself — the seismic reading, the alert text, the filing — which is upstream of any outlet’s write-up of it. Most aggregators can only reach the first, because their input is other people’s articles. An aggregator that watches agency feeds directly can reach the second for the categories those feeds cover, and cannot reach it for anything else. Neither is better in the abstract. What matters is that a product tells you which one it is doing.
The Front Wire: the source is the destination, not the exhibit
The Front Wire watches four kinds of feed — the US Geological Survey earthquake feed, National Weather Service severe-weather alerts, SEC EDGAR filings and the major newswires — and every item on the wire carries the outlet or agency it came from. Be precise about the hops, because this is where aggregators usually get vague, and being precise means admitting the two surfaces differ. On the site, a headline opens a Front Wire page carrying the item’s own details, and that page ends in a link out to the original report on the source’s site. In the daily digest email there is no intermediate page at all: every headline links straight to the publisher or the agency. In the RSS feed there is one — each item’s <link> is the Front Wire page, and the publisher’s own URL travels beside it in that item’s <source url> attribute. So: none in the email, one in the feed, one on the site. For seismic, weather and filings that destination is the agency record rather than another outlet’s summary of it, which is the second sense of primary source above. Nothing is personalised: two readers loading the wire at the same second see the same items in the same order. Reading it is free and needs no account; the membership at $3/mo or $24/yr pays for running the wire and changes nothing about what a reader sees or where a link points. Membership prices read from the live Stripe prices behind the checkout (fw_monthly, fw_yearly), 3 August 2026. Hop behaviour re-read off the live feed.xml and the digest sender, 12 August 2026 — the feed sentence above was wrong until then and has been corrected. Feeds are first-party — see our sources.
Techmeme: human editors, tech only
Techmeme is the strongest answer for one reader in particular: someone who only wants technology news and wants a person, not a model, deciding what leads. Its About page describes an editorial pyramid where “editors make final calls on what we feature, and write descriptive, straightforward headlines for that news”, with automated crawling and filtering underneath rather than on top. A team of editors “across five continents work around the clock”, sourcing “from thousands of outlets”, and every headline carries the outlet’s name beside it. The site has “remained self-funded and independently-owned since its 2005 launch”, supported by advertisers, data sales and news-filtering services rather than reader subscriptions. Its limit is its scope: it is a tech aggregator, so an earthquake, a tornado warning or an 8-K from a non-tech filer is not something it is trying to cover. If tech is what you read, this is the better product, and that is not a close call. Source: About Techmeme. Checked 3 August 2026.
Ground News: the same event, across outlets
Ground News solves a different problem from routing: it is built to show you how the same story is being framed by different newsrooms. Its About page states that “every day we process nearly 60,000 news articles from over 50,000 different news sources”, and that “articles from different outlets covering the same event are merged into a single story, making it possible to get multiple perspectives in one place” — the product describes itself as a way to “compare news sources, read between the lines of media bias and break free from algorithms”. You still reach a publisher, but you choose which one from the merged list, which is a genuinely useful extra step when coverage is contested and pure overhead when it is not. A magnitude reading is not framed by anybody. If what you want is to see who is saying what about a political story, this is a better fit than any wire, ours included. Source: About Us — Ground News. Checked 27 August 2026.
Readless: a summary layer over the sources you already chose
Readless answers this question differently enough from the rest of the table to be worth naming, because it is not reading the open web at all. Its inputs are yours: you forward newsletters to an “@mail.readless.app address”, “Paste any Substack handle, profile URL, or publication URL”, or “Drop any RSS feed URL into a digest schedule”, and it turns “30+ newsletters, Substack publications, and RSS feeds into one 5-minute AI digest”, “deduplicating overlapping stories and stripping ads before delivery”. On this page’s axis it does link out, and says so plainly: “Every summary includes a direct link to the original article.” What sits between a reader and that link is a model’s summary rather than a headline — “a GPT-5-level model tuned to capture the most important information while maintaining context” — which is a difference in kind rather than a fault, since reading less is the stated point. Two consequences follow for anyone comparing it with a wire. Nothing reaches you that you did not already subscribe to, so a magnitude 6.1 or a tornado warning arrives only if some newsletter you added happens to cover it. And the first thing you read is generated text about the reporting rather than the reporting. Pricing is $4.90/mo for Pro, with up to three digest schedules and unlimited sources, and $9/mo for Max, with up to eight schedules and a cross-source editorial briefing; a seven-day Pro trial requires a card. Sources: Readless, How it works and Pricing. All checked 27 August 2026.
Google News: the widest net, and an algorithm that knows you
Google News has coverage no independent aggregator can match, and it is honest in its own documentation about the trade. Google’s help page on how stories are selected states plainly: “Computer algorithms determine which stories show up on Google News and in what order”, and “Google News shows some content in a personalized way”. The personalisation inputs are named — your Google News settings, and “past activity: your activity on Google services and YouTube” — across the For you, Following, Your topics, Local news and Notifications sections. On the publisher side, Google’s Publisher Center states that “Google may find your site through our normal web crawl”, and that blocking Googlebot-News “excludes your content from Google News” entirely. Two consequences follow for anyone comparing aggregators. The set of stories you see is not the set your colleague sees, so nobody can tell you what was on Google News this morning. And a source only appears if it has been crawled and deemed eligible, which is a different mechanism from reading a feed the moment it publishes. No fee is named anywhere in Google’s help pages for either. Sources: How Google News stories are selected and News content across Google. Both checked 3 August 2026.
The shortest path is no aggregator at all
For three of the four source types on this page, you can skip every product listed and read the primary record directly, for free and with no account. The USGS publishes its earthquake catalogue as public GeoJSON summary feeds, the National Weather Service serves US severe-weather alerts from api.weather.gov, and SEC EDGAR publishes an Atom feed of the newest filings filterable by form type. There is no aggregator between you and any of them, which makes this the honest baseline any aggregator has to beat. What you take on instead is the polling, the User-Agent handling, the CAP and Atom parsing, and the fact that four separate feeds do not sort themselves into one list — every day, including the days nothing happens. Every feed we watch is named, with its endpoint and the date it was last checked, on our sources page, so you can take any of them and leave the wire behind. The two hazard feeds get their own page: a news feed of USGS earthquake and NWS weather alerts sets the raw GeoJSON and CAP against the notification services each agency runs itself. Feed details, fee and User-Agent terms carried from the checks recorded on our sources page, 2 August 2026.
So: which one
Technology news, chosen by editors — Techmeme, and it is not close. How a contested story is being framed across newsrooms — Ground News. A pile of newsletters you already subscribed to, cut down to one summarised digest — Readless. The broadest possible coverage, accepting that the selection is algorithmic and partly tuned to you — Google News. One feed and total control, if you are willing to write the polling — go straight to the USGS, the NWS and SEC EDGAR. Earthquakes, severe weather, filings and newswire headlines in one list, ranked the same for everyone, with every item linked to the outlet or agency it came from — that one is ours, and the free daily digest below is the least effortful way to try it. If your question is really about monitoring a keyword rather than reading a wire, the comparison you want is alternatives to Google News alerts; if it is about news from before any of this existed, nothing here covers that, because the historical newspaper archives are a different product and this site does not compare them; and if what you want is one email rather than a site to visit at all, breaking news digest email subscriptions compares the digest below against the paid ones.
Subscribe to the daily digest
Free. One email a day with the day’s eight top items — every headline linked straight to the outlet or agency it came from, with no page of ours in between.
Back to the wire, or open the archive.