MORNING MOSAIC Get the brief

One clear picture of the world, before your day starts.

Morning Mosaic reads several hundred stories from 28 news sources every morning and assembles them into a single short brief. Plain language instead of hyperbole. Every outlet's lean printed next to its name, so the source and where it generally sits are visible together. Ranked by what actually affects people, and it ends with the genuine progress that most of the news day leaves on the floor.

Arrives 6:45am Eastern Twenty-eight sources, read this morning Every claim links to the original
The top of a real Morning Mosaic brief: the wordmark, the date Thursday July 30 2026, an opening line, and the first two ranked stories, each with a plain summary, a why-it-matters line, an across-the-spectrum note, and source links labelled Center, Left, Right and Center-Left.
A real brief, sent on 30 July 2026. Not a mockup: this is the file that landed in an inbox that morning.
Why this exists

You did not stop caring. You stopped being able to stand it.

Plenty of people have stopped opening the news, and the reasons they give are not that the world got boring. They report that opening a news app in the morning has become an unpleasant thing to do to yourself, and because it is genuinely hard to tell what is true, what is being shaped, and what is simply being shouted the loudest for the largest number of seconds.

So the day starts one of two ways. Either you skip it, and carry a low hum of guilt about being a person who no longer knows what is going on. Or you do read it, and you walk into your morning holding 40 fragments you cannot order, half of which will turn out to have been the same story wearing four different headlines. Neither of those is being informed. They are just different ways of losing the same 20 minutes.

The survey figures below say this is now common rather than unusual. They measure how many people avoid the news and how much they trust it; the reading that follows, about an industry paid by the second and a reader left to do the sorting, is ours rather than the report's.

News avoidance, worldwide

42%

Of people who say they actively avoid the news at least sometimes, up from 29% a decade ago. In the United States the figure is 45%.

Reuters Institute Digital News Report 2026

Trust in news, across markets

37%

Close to the lowest level the same survey has recorded since it began tracking it.

Reuters Institute Digital News Report 2026

Would trust a chatbot for news

20%

One reader in five. That is why this product leads with its sources and its method rather than with the machinery.

Reuters Institute Digital News Report 2026

Morning Mosaic is built on a plain wager: most people do not want less news, and they do not want a machine's opinion of the news. They want the same information, assembled once, in an order that reflects what it will actually do to people, in language that does not spend their attention before breakfast. And they want to be able to check it, which is why every line in the brief carries a link back to the reporting it came from.

The three sections

Three parts, in the same order, every morning.

The order does not change, which is the point. After a week you stop reading the brief and start scanning it, because you already know where everything is. It takes about four minutes to read properly and about 90 seconds to skim.

The Briefing

The stories that matter most today, in order. Each one gets what happened in plain sentences, a short line on why it matters to anyone who is not already following it, and, when coverage genuinely splits, a note on how different parts of the press are framing it. Then the sources, each labelled.

The Quick Mosaic

Everything else worth a sentence. One line, one link, no ceremony. This is the part that catches the story you would otherwise have found out about a week late from somebody at dinner.

Bright Tiles

Real progress, reported as news rather than as a consolation prize at the bottom. A treatment that worked, a number that moved the right way, a conflict that de-escalated. It is in the brief because it is true, not to make you feel better.

Read the whole thing.

This is the complete brief from 30 July 2026, top to bottom, at the size it arrives. Look at the source lines under each story: BBC News (Center) · The Guardian (Left) · Fox News World (Right) · Al Jazeera (Center-Left). That label describes the outlet's general lean, never the individual story, and it is applied from our own published source table rather than being invented per article.

You will also notice what is missing. No image blocks that fail to load, no tracking pixels dressed as pictures, no sponsored slot in the third position. The brief is typographic on purpose, so it stays readable with images switched off, in dark mode, and in the mail clients we test.

Every headline links out. We summarise and we point at the original; we never reproduce an article's text. If a story matters to you, the point of the brief is to get you to the reporting, not to replace it.

A full Morning Mosaic brief showing six ranked stories with summaries, why-it-matters lines and labelled source links, followed by the Quick Mosaic and Bright Tiles sections.
The 30 July 2026 edition, unedited.
How it works

Six steps, finished before the 6:45am send.

There is no mystery here and no black box worth protecting. This is the whole pipeline, in the order it runs.

  1. Read everything

    Every source on the published list is fetched directly from its own feed, in parallel, starting a few minutes before the brief is due. This morning that produced several hundred fresh items. If fewer than 50 come back, the run stops and no brief is sent, because a thin brief is worse than a missing one.

  2. Rank by consequence, not by volume

    The ordering question is how many people this changes something for, and how much. That is deliberately not the same as how many outlets are covering it, which is the metric that produces a front page of one story in nine costumes.

  3. Rewrite in plain language

    Each selected story is written out in ordinary sentences: what happened, and one line on why it matters. Hyperbole, unattributed drama, and the words that exist to make you feel something before you have understood anything are removed rather than softened.

  4. Label the sources

    Leans come from our own source table, which is published below and applied to the outlet, never to the story. A label tells you where a publication generally sits. It does not tell you that a particular piece is wrong, and it is not a score.

  5. Check it against the pool

    Before anything is sent, a validator re-derives every lean from the source table and drops any citation whose link was not in the fetched pool. On the very first run it caught an invented URL and removed it. This step is not optional and it is not tuned down.

  6. Send at 6:45am Eastern

    Rendered as a typographic email, contrast-measured against the accessibility standard by a test that ships with a deliberate failing case so the test itself can be trusted, and delivered from our own domain.

Sources and method

The part most briefings will not show you.

Only one in five people say they would trust a chatbot to tell them what happened today, and that is a reasonable position to hold. The answer is not to argue about it. It is to publish the method, so you can check the thing instead of trusting it.

Where the news comes from

Twenty-eight feeds today, read directly from the publications themselves, spanning left to right and including international desks. Among them: the BBC, The Guardian, NPR, The New York Times, The Wall Street Journal, Fox News, Al Jazeera, CBS News, ABC News, NBC News, CNN, CNBC, Politico, Axios, The Economist, National Review, Vox, Sky News, France 24, Deutsche Welle, and UN News. The whole table is printed below, with the lean assigned to each one. The list grows, and it is versioned, so a change to it is a visible change rather than a quiet one.

What a lean label means

Five positions: Left, Center-Left, Center, Center-Right, Right. It is an editorial judgment about the publication, held in one file, applied consistently, and never generated per story by a model. If you disagree with a label, you are disagreeing with a specific published line rather than with a hidden process, which is the entire reason it works this way.

What we never do

We do not reproduce article text. We do not paraphrase closely enough to be a substitute for reading the source. We do not build on aggregator feeds that repackage other people's reporting without permission. Summaries are short and genuinely transformed, and the link out is prominent by design, not buried in a footer.

What happens when it is wrong

Any automated writing step can produce a claim nobody made. Rather than promising that ours will not, we assume it will and check: every cited link must already exist in the material fetched that morning, or it is removed before sending. That is a mechanical guarantee about citations. It is not a guarantee that a summary is never clumsy, and we will not pretend otherwise.

The source table

All 28 feeds, with the lean we assign each one.

This is the file the brief is built from, printed in full. It is here because the rest of this page keeps claiming the labels are checkable, and a claim like that is worth nothing without the table beside it. Disagree with a row and you are disagreeing with a specific published line, which is the entire point.

The spread today: three Left, nine Center-Left, 10 Center, two Center-Right, four Right. That is not a balanced set and we are not going to describe it as one. It leans to the centre and centre-left, mostly because that is where the large general-news desks with open feeds sit, and widening it is ongoing work rather than a finished job.

Every publication Morning Mosaic reads, with its assigned editorial lean
PublicationAssigned leanDesk
BBC NewsCenterInternational
BBC Top StoriesCenterInternational
NPRCenter-LeftUnited States
The GuardianLeftInternational
The New York TimesCenter-LeftUnited States
NYT Front PageCenter-LeftUnited States
The Washington PostCenter-LeftUnited States
The Wall Street JournalCenter-RightUnited States
Fox NewsRightUnited States
Fox News WorldRightUnited States
Al JazeeraCenter-LeftInternational
CNBCCenterUnited States
PoliticoCenter-LeftUnited States
The HillCenterUnited States
AxiosCenterUnited States
CBS NewsCenter-LeftUnited States
ABC NewsCenterUnited States
NBC NewsCenter-LeftUnited States
Sky NewsCenterInternational
France 24CenterInternational
The EconomistCenter-RightInternational
National ReviewRightUnited States
New York PostRightUnited States
VoxLeftUnited States
Mother JonesLeftUnited States
CNNCenter-LeftUnited States
Deutsche WelleCenterInternational
UN NewsCenterInternational

Three publications appear twice because two of their feeds are read separately: the BBC, The New York Times, and Fox News. That is why the table has 28 rows and 25 distinct publications. A label describes the publication, never the individual story, and it is not a factuality score: a Right-labelled outlet breaking an accurate story is an accurate story, and the label is there so you know whose desk it came from.

What is built, and what is not

Nothing on this page is sold as finished unless it is.

The daily email works today and has been sending every morning since 30 July 2026. Everything else on the road map is marked as planned, and planned means not built. If a row says planned, do not pay for it and do not plan around it.

Morning Mosaic capabilities and their build status
CapabilityStatusWhat it means
Daily email briefBuilt Sends at 6:45am Eastern from our own domain, on any morning the source pool clears the refusal floor.
Outlet lean labelsBuilt Five-point scale from a published source table, applied to publications.
Impact rankingBuilt Ordered by consequence rather than by how many outlets ran it.
Progress reported as newsBuilt The Bright Tiles section, in every edition.
Citation validationBuilt A cited link that was not in the morning's fetched pool is dropped before sending.
Across-the-spectrum noteBuilt How coverage differs, on the stories where it genuinely does.
Renders with images offBuilt Typographic by design, contrast-measured in a test with a failing control case.
Licensed wire copyPlanned An application to a wire service's developer programme, not yet submitted.
Third-party bias and tone dataPlanned Would supplement our own labels with an independent provider's. Not wired in.
Daily text message editionPlanned Requires carrier registration in the United States. Nothing is built.
Choose your own topicsPlanned Everyone receives the same brief today. There is no personalisation at all.
Searchable archive on the webPlanned Past editions are not published anywhere yet.
Evening editionPlanned A second send on days the story moves. Not built.
iPhone appPlanned Not started. Email is the product today.
The moments it is for

Four minutes, at four different kitchen tables.

The 6:50am scroll, replaced

You reach for your phone before your feet hit the floor, and 20 minutes later you are angry about something that will not affect you and still do not know what happened in the world. One email, read while the coffee brews, ends that particular loop.

Walking into a room informed

Client breakfasts, school runs, standing meetings where somebody will mention the thing that happened overnight. Knowing the five things that actually moved, and where each one was reported, is worth more than having read 40 headlines.

When you have opted out entirely

If you deleted the apps and turned off the alerts and it genuinely helped, this is the version that does not undo that. It arrives once, it has an end, and it never buzzes.

Reading across the line

If you already read widely and want to see, on a given story, that the right-leaning coverage led with one fact and the left-leaning coverage led with another, the across-the-spectrum note and the labelled links put both in front of you in one place.

Compared with

The others, honestly.

Every price below was read from the vendor's own page on 30 July 2026 and linked, and where a vendor's pricing page could not be read, this table says so rather than repeating a number from somewhere else. The rows where somebody else wins are marked as their win, because a comparison that never loses is a comparison nobody should believe.

Morning Mosaic compared with other daily news briefings
ProductPrice, from their page Outlet lean labelledProgress as a section Where it is ahead
Morning Mosaic Free todayNo paid tier is live. Proposal below. Yes, on every source link Yes, in every edition One assembled brief with the lean printed inline and a citation check that removes links it cannot find in that morning's material.
Ground News $8.33/month, billed annuallyVantage plan, plus a free tier Yes, and factuality ratings too No Ahead of us: far more sources, licensed bias and factuality ratings from independent raters, and a shipped app and browser extension. We have none of those.
Tangle $6/month or $59/yearBundle with podcast $10/month or $99/year Argument by side, not a per-outlet label No Ahead of us: a named human editor taking one story a day and laying out the best arguments from left and right in depth, with a podcast. That is journalism, and we do not do it.
1440 Free"100% free, unsubscribe anytime" No Occasionally, not as a section Ahead of us: millions of readers and years of consecutive daily editions. Our track record is measured in days.
News Minimalist Not readableTheir pricing page is behind a bot check, so we are not quoting a figure No No Ahead of us: a live web feed you can browse whenever you want, ranked continuously rather than once a day.
Boring Report Not readableNo price stated on the page we could read No No Ahead of us: a real mobile app, and a longer history of stripping sensational language from stories.

Read plainly, that table says Ground News is the more complete product, Tangle is the better piece of writing, and 1440 has the audience. What none of them does is the specific combination this one is built around: a single assembled brief where the lean sits next to every link, the ordering is by consequence, the progress is treated as news rather than as a footnote, and the citations are mechanically checked before anything is sent.

Pricing

The email is free. The rest is a proposal.

Draft, not final

There is no way to pay for anything on this site, no card form, and no checkout. The numbers below are a proposal being worked through, published here so it is visible rather than decided quietly. They will move.

The daily brief

Free
and intended to stay that way
  • The full brief, every morning at 6:45am Eastern
  • All three sections, all sources labelled
  • Every claim linked to the original reporting
  • No advertising in the brief

Pro Planned

$14/month
or $120 a year, proposed, nothing is charged
  • The same brief as a morning text message
  • Weight the ranking toward the subjects you actually follow
  • A searchable archive of every past edition
  • An evening edition on the days the story moves

Every one of those four is unbuilt. This tier does not exist and cannot be bought.

The reasoning, in plain terms: the daily email is the thing people will forward, so charging for it would strangle the only channel that grows this. The proposed $14 sits above Ground News at $8.33 a month billed annually and above Tangle's $10 bundle, which is deliberate. A briefing that saves a real 20 minutes a day is not competing on being the cheapest, and a price set low is very hard to raise later. The paid tier has to earn that number, and today it is a list of four unbuilt things, which is exactly why nothing is being charged yet.

Privacy

What we hold, which is almost nothing.

To send you an email we need your email address. That is the whole list of personal information involved, and there is no second use for it.

  • Your address is never sold, rented, or shared with another company. There is no data partner, no list broker, and no arrangement under which somebody else mails you because you signed up here.
  • The brief carries no tracking pixel. It is a typographic email with no images at all, so there is nothing embedded in it that reports back when you open it. We do not know who reads it.
  • Leaving takes one click and is immediate. Every edition carries an unsubscribe link, and using it removes the address rather than flagging it.
  • This site measures traffic in aggregate. Page views and which sections get read, with text and form inputs masked, so an email address typed into the box above is never captured. Do Not Track is respected: if your browser sends it, the measurement does not run at all.
  • The confirmation step is deliberate. Nothing is sent to an address until somebody has confirmed it from that inbox, so a stranger cannot sign you up.
Limits

What it does not do, stated plainly.

This section exists because a page that only lists strengths is not describing a product, it is describing a wish. Every item here is true today.

It is young

The first edition was sent on 30 July 2026. The list is small and is being opened up as the sending infrastructure is finished, so a confirmed address may wait before the first brief arrives. There is no queue-jumping and nothing to pay to move up it.

It is not personalised

Everyone gets the same brief. You cannot yet exclude a subject, follow a country, or change the length. That is on the road map above and it is not built.

It is English-language and Anglosphere-weighted

The source list leans towards United States, United Kingdom, and international English desks. It carries international coverage but it is not yet a genuinely global read, and calling it one would be a claim we cannot support.

Lean labels are a judgment

They are ours, they are published, and reasonable people will disagree with some of them. They describe outlets and not stories, and they should be read as a signpost rather than as a verdict on any individual piece.

It summarises, it does not report

Nobody here is on the ground. The value is in the selection, the ordering, and the plain rewriting, and the links exist so you can go and read the people who did the actual work.

Automated writing can still be clumsy

Citations are checked mechanically, so a link that was not in the morning's material never ships. Wording is not mechanically checked. A summary can be flat or can miss a nuance, and the honest answer is to read the source.

FAQ

Frequently asked questions

What does it cost?

The daily email is free and is meant to stay free. There is no card form on this site and nothing to buy. A paid tier is proposed at $14 a month in the pricing section above, and every capability listed under it is unbuilt, so it cannot be purchased.

When does it arrive?

6:45am Eastern, on any morning there is enough news in the pool to brief. The pipeline starts a few minutes earlier and the whole run, from fetching feeds to sending, finishes in that window. If fewer than 50 fresh items come back, nothing is sent that day, because a thin brief is worse than a missing one.

How many sources does it read?

Twenty-eight feeds, from 25 publications, read straight from their own sites. Three publications contribute two feeds each. The most recent run produced several hundred fresh items. If fewer than 50 fresh items come back, no brief is sent at all.

What do the labels next to each outlet mean?

They are that publication's general editorial lean on a five-point scale from Left to Right. They come from our own source table, printed in full on this page. They are applied to the outlet and never to an individual story, and they are not generated per article by a model.

Is this written by a machine?

The summarising and ordering are automated. The source list, the lean labels and the editorial rules the brief is written to are human decisions, held in files, and published. Only one in five people say they would trust a chatbot for news, which is why the method is on this page and every claim in the brief links back to the reporting it came from.

Does it reproduce articles?

No. Summaries are short and genuinely transformed, the link out to the original is prominent by design, and full text is never reproduced. If a story matters to you, the brief is built to send you to the people who reported it.

What stops it inventing a source?

A validator runs before anything is sent. Every cited link must already exist in the material fetched that morning or it is removed, and every lean is re-derived from the source table rather than taken from the generated text. On the first run it caught and removed an invented link.

How do I unsubscribe?

Every edition carries an unsubscribe link that works in one click and removes the address. You can also reply to any brief and ask.

Will you sell my email address?

No. It is used to send you the brief and for nothing else. It is not sold, rented, or shared with another company, and the brief carries no tracking pixel, so we do not know who opens it.

Why is there a section of good news?

Because leaving it out is itself a distortion. Treatments work, conflicts de-escalate, and numbers move the right way, and a briefing that only reports the other kind gives a false picture of the day. Bright Tiles is in every edition and it is reported to the same standard as everything else.

Is there an app?

Not yet. Email is the product today. A text-message edition and an iPhone app are both on the road map and both are marked planned, which means not built.

Who is behind it?

Morning Mosaic is built by Belstone. It has been sending every morning since 30 July 2026.

Wake up with the whole picture.

One email a day, free, four minutes, and every line of it links back to the reporting it came from. The list is small while the sending rail is being finished, so a confirmed address may wait a few mornings for its first brief.