About
The world, from its sources.
browse.info shows what is happening in the world as the people who measure it report it: earthquakes, disaster and weather alerts, storms, the news as Wikipedia's editors, governments and the registered publishers' own feeds report it, exchange rates, space weather, the space station, and countries' published figures. It is free, it has no accounts and no advertising, and it writes none of the news.
Who makes it
Browse is made by Yank, the site for keeping sourced findings as cards with their receipts. Write to comboventures+browse@gmail.com.
How to read a figure
Every figure on Browse comes with three things: its source, named and linked; its time, the source's own, in UTC or with its time zone; and its kind, so a measurement is never mistaken for an alert, a forecast or Browse's own arithmetic.
- OBSERVED
- The source reports it happened or measured it.
- ALERT
- An authority says it may happen or is expected; an alert is not the event.
- FORECAST
- A prediction, by the source that made it.
- SCHEDULED
- A time the source has published for something planned, which may still change.
- OFFICIAL RECORD
- A government's own publication.
- STATISTIC
- An official figure the source published for a date.
- ESTIMATE
- The source's own model estimate, not a count.
- REFERENCE
- An encyclopedia entry or sourced factual page.
- MEDIA
- A picture or video, with its author and licence.
- DERIVED
- Browse's own calculation from the records named, by the operation shown.
Sources & methods
Where every figure on Browse comes from, below, and how Browse works out the ones it computes, in How the figures are worked out and How News is made.
Earth and weather
- USGS Earthquake Hazards Program: earthquakes of magnitude 4.5 and above, and significant ones (public domain).
- GDACS (UN and European Commission): Orange and Red disaster alerts worldwide (CC BY 4.0).
- US National Weather Service: active US alerts rated Severe or Extreme, local alerts for a place, and the forecast for US places (public domain).
- NOAA National Hurricane Center: active tropical storms and their advisories.
- MET Norway (Norwegian Meteorological Institute): the forecast for places outside the United States (CC BY 4.0).
- Aviation Weather Center: observed weather at the station nearest a place (METAR; public domain).
News and records
- Wikipedia and Wikimedia: In the news, Current events, featured content, page views, On this day, the Deaths in the year list, and freely licensed pictures from Commons.
- The White House, the Federal Register, the European Commission, the UK Government (gov.uk): governments' own records and statements, by their own titles and times (US government work, CC BY 4.0, Open Government Licence v3.0).
- SEC EDGAR: the newest current reports (8-K) companies file, by form, company, the items the filing lists and its accession number (US government work; read with Browse's named User-Agent, as the SEC's fair-access rules ask).
- Voice of America: its newsroom's headlines, authors and times, each linked to the original (VOA's own work is public domain; nothing of the text is shown).
- Global Voices: its newsroom's headlines, authors and times, each linked to the original (CC BY 3.0).
- The Conversation: the headlines, authors and times of its US, global, UK, Australian, Canadian and African editions, each linked to the original (CC BY-ND 4.0).
- Inside Climate News: its newsroom's headlines, authors and times, each linked to the original (CC BY-NC-ND 3.0; Browse is a non-commercial use: no advertising, no sales, no accounts).
- ProPublica: its newsroom's headlines, authors and times, each linked to the original (CC BY-NC-ND 4.0; Browse is a non-commercial use: no advertising, no sales, no accounts).
- UN News: the United Nations' own news service's headlines and times, each linked to the original (the UN's terms allow personal, non-commercial use; Browse is a non-commercial use: no advertising, no sales, no accounts).
- The publishers' own feeds (Coverage): the public RSS, Atom or news-sitemap feeds of the publishers on the register, read the way a feed reader does with a reader that names itself and stops where robots.txt says so: the link, the headline as the outlet wrote it, the time, the author, the categories and the address of its picture, never the text; grouped by a printed rule, each story headed by its earliest outlet's own words; an outlet is removed within 48 hours of asking.
- The GDELT Project: its newest fifteen-minute Global Knowledge Graph files, the English stream and the stream of 65 other languages: every news page its crawlers read, by the page's own title and time and GDELT's name for the site, with the places GDELT read in it (unrestricted use, cited and linked).
Money and figures
- European Central Bank: daily euro foreign exchange reference rates, set around 16:00 CET on working days and not live; and, in News, its press releases, speeches, interviews and monetary policy accounts by their own titles and times (reproduction permitted with the ECB cited).
- World Bank: countries' published indicators: population, births, deaths, life expectancy and more (CC BY 4.0).
- US Treasury Fiscal Data: the US national debt, as the Treasury records it each working day (public domain).
Space
- NASA eclipse predictions: the solar and lunar eclipses of 2026 to 2030 on /space's sky calendar (Eclipse Predictions by Fred Espenak, NASA's GSFC).
- NOAA Space Weather Prediction Center: geomagnetic storms, the Kp index, alerts and forecasts, the OVATION model's chance of an aurora overhead, and the GOES-19 CCOR-1 coronagraph's pictures of the corona (public domain).
- Launch Library 2 (The Space Devs): upcoming launches and the crew in orbit.
- CelesTrak (and ARISS when CelesTrak cannot be read): the space station's orbital elements, which the US Space Force's 18th Space Defense Squadron publishes on Space-Track.org; Browse computes the station's position and passes from them, and names whichever source gave them.
- NASA/JPL Center for Near-Earth Object Studies: asteroid close approaches.
- NASA Solar Dynamics Observatory (SDO): the Sun in AIA's ten channels and HMI's intensitygram and magnetogram, every ten to fifteen minutes since 2016, from SDO's browse pictures (courtesy of NASA/SDO and the AIA, EVE, and HMI science teams).
- NASA DONKI (Community Coordinated Modeling Center): solar flares, coronal mass ejections with their analysts' measurements and model runs, geomagnetic storms and the analysts' notifications, as DONKI records them (prototyping quality, in CCMC's words).
- NASA EPIC (on NOAA's DSCOVR): the whole sunlit Earth from the Sun-Earth L1 point, a day at a time since June 2015, in natural and enhanced colour (free to reuse, credited to the NASA EPIC Team).
- NASA 3D Resources: NASA's own 3D models of spacecraft and an asteroid, shown in Browse's viewer from Browse's copy (decoded and made smaller for the web; shapes and colours NASA's).
- JPL Solar System Dynamics: approximate planetary positions: the planets' orbital elements and their rates (Standish, valid 1800 to 2050), from which Browse computes where the planets are at any moment.
- NASA's Eyes (NASA/JPL): NASA's own 3D apps (the solar system, the Earth, asteroids, exoplanets, the Deep Space Network), loaded inside Space in 3D only when a reader taps, under NASA's own privacy policy.
- NASA GIBS, Blue Marble and Black Marble: the globe's picture of the Earth: yesterday's satellite view, and the night lights.
Cameras
- Caltrans (California Department of Transportation): the stills of California's highway cameras, for Live views.
- USGS Hawaiian Volcano Observatory: the stills of Kīlauea's summit cameras, for Live views (public domain).
- Windy.com Webcams API: public webcams worldwide for Live views, each published by its owner, shown under Windy's terms with its credit and each webcam's timelapse.
Places
- date-holidays: public holidays for a followed place's country (CC BY-SA 3.0 data, ISC code).
- IANA time zone database: a place's next clock change, read through the browser's own clock rules.
- adsb.lol: the aircraft its volunteer receivers report over a place, on Planes overhead (ODbL 1.0).
- GeoNames: the towns and cities of 5,000 people or more behind place suggestions as you type (CC BY 4.0).
- US Census Bureau, ZIP Code Tabulation Areas: the US ZIP codes behind place suggestions as you type (2024 gazetteer, public domain).
- OpenStreetMap Nominatim: finding a place by its name.
History and reference
- Wiktionary: the word of the day in Discover, in Wiktionary's own words (CC BY-SA 4.0).
- Library of Congress, Chronicling America: the front pages of American newspapers from a hundred years ago (public domain).
- The Nobel Prize: the laureates of each year's prizes as they are announced (CC0).
- Wikidata: the photograph its editors chose for a place; how many Wikipedia languages have an article on a person (the count that ranks On this day); films' and games' release dates.
Culture and sport
- TVmaze: TV schedules by country and new episodes on streaming services (CC BY-SA).
- Wikipedia's season articles: league tables as its editors record them, with their updated-to date and source (CC BY-SA 4.0).
How the figures are worked out
- Times
- A source's time says when it published or revised something. "Updated" says when Browse last read the sources; an open page asks again every minute. "Ago" is counted from your clock.
- Reach
- Severe-weather alerts cover the United States only, because they are the National Weather Service's. The other event feeds are worldwide.
- Live counters
- The population, birth and death counters are estimates, never a count: they start from a published population figure (the World Bank's mid-year estimate, or the US Census Bureau's clock when it can be read) and move by the World Bank's published birth and death rates.
- Weather beside a place
- The nearest airport's latest observation. A place's sky picture follows the Sun's height there, worked out on the device; it is not a photo or a forecast.
- What changed
- Each change a check finds is dated in Browse's own record, so a row can say what changed between two updates.
- When a source cannot be read
- The page says so where that source's information would be, with its last good update when Browse has one. Nothing is filled in.
How News is made
The News tab, step by step, written from the code it runs on. Every story page carries the same in short, with a report link.
- What is read
- Wikipedia's In the news (the Main Page template and its revision history) and Current events (one page per day, read every five minutes for the last three days and hourly for older days)
- The White House, Federal Register, European Commission, UK Government and European Central Bank feeds and the SEC's EDGAR feed of current 8-K filings
- The feeds of six newsrooms whose own licences or terms allow reuse, Voice of America (public domain), Global Voices (CC BY 3.0), The Conversation (CC BY-ND 4.0, six editions), ProPublica (CC BY-NC-ND 4.0), UN News (personal, non-commercial use) and Inside Climate News (CC BY-NC-ND 3.0), the last three for non-commercial use, which Browse is (no advertising, no sales, no accounts), of which only the headline, author, time and link are shown
- The GDELT Project's newest fifteen-minute Global Knowledge Graph files, the English stream and, when a reader picks a language, the stream of 65 other languages, every news page its crawlers read, under GDELT's terms (unrestricted use, cited and linked), of which only the page's own title in its own language, its address, its time, GDELT's name for the site and the places GDELT read are shown.
- Since 19 October 2026 the public feeds of the publishers listed at /news/publishers (26 at first, 69 from 22 October, 70 with the Mirror from 5 November, across the United States, the UK and Ireland, Europe, the Middle East, Africa, Asia, Oceania and the rest of the Americas, with science, technology, business and sports outlets; each measured one by one: the feed, its dates, its robots.txt), read the way a feed reader does, by a reader that names itself, every ten minutes at most, stopping where robots.txt says so; of each item only the link, the headline as the outlet wrote it, the time, the author and the categories are kept, grouped into stories by one printed rule with the earliest outlet's own headline as the heading and outlets counted once.
- Several of those outlets' terms speak of personal or non-commercial use; Browse reads them as a non-commercial use (no advertising, no sales, no accounts) on the owner's decision of 19 October 2026, quotes each outlet's own sentence on the register, and removes an outlet within 48 hours of its written request.
- A page found through GDELT is shown by its title and address on the owner's decision of 11 October 2026, and a publisher that asks in writing not to be listed is removed, with the date (none has asked).
- Since 31 October 2026, on the owner's decision, a story in Coverage or the brief can show one outlet's own share picture: the heading's outlet's or, if it has none, the earliest other outlet's that has one. The address comes from the feed entry or, for a story's lead item with none, once from the article page's og:image or twitter:image (robots.txt first, Browse's named reader, the page's text never kept). The reader's browser loads it directly from the outlet with no referrer, credited to the outlet and linked to its article; Browse never copies, stores or resizes it, and the free picture stands in when there is none, when one fails to load, or when the reader turns them off. An address on several of an outlet's articles is its default image and is left out. Once a week at most one recent article of each outlet is read for its robots rules: max-image-preview:none, noimageindex or noindex on images turns that outlet's pictures off, with the date on /news/publishers, as does the outlet's written request within 48 hours. Share cards, the map, Discover, Browse's feeds and these agent files never carry a publisher's picture.
- Since 31 October 2026, also on the owner's decision, the home page shows the most carried stories' outlets' own pictures, under the same rules and the same device switch, each with that outlet's own headline. The largest usable picture wins: of a story's outlets, one at least 800 pixels wide before one of no stated size before a smaller one; a feed's small copy gives way to the article's own share picture; a tiny, logo-like, banner-shaped or very tall file is left out. Beside an outlet's name is its small icon, read once from its own site by a script run by hand and kept as a small copy on Browse's server, so no page asks the outlet for it; an outlet with none gets a monogram, and an icon comes off within 48 hours of the outlet's request, with the date in the terms register. An outlet's name leads to a page of its current headlines at /news/outlet/<key>.
- Not read: the text or summary of any article, and publishers not on the register at /news/publishers.
- What is kept
- Nothing of Browse's own: each line is served from Wikipedia's current revision, cached a few minutes, with the revision number as its receipt. A cited page is read only when a reader asks (the quote check), and then only its verdicts are kept for a day, never its text. Of each registered publisher's item, only its link, headline, time, author, categories and its picture's address and alt text are kept, as said above; never the picture. Each records feed and newsroom feed keeps its last good read while its source cannot be read or answers with nothing, and prints that read's time beside it.
- How lines are grouped
- A story on Wikipedia's side is Wikipedia's own heading: the article its editors link above a line. Browse names nothing there, and a line under no heading stands alone. The publishers' own headlines are grouped into stories by the rule Coverage prints beside them, labelled DERIVED: A headline joins a story when it shares three or more words of four letters or more (names, places, things; common words set aside) with the story's first headline and was published within 48 hours of it. Unless four or more are shared, the shared words must include a name, or two words that stand side by side in both. A name is a word both headlines capitalise that the read's headlines mostly write with a capital away from their first word (Title Case headlines aside), carried by no more than one in a hundred of the stories (or six, when that is more); two capitalised words side by side in both, such as White House, are one name and count as one word. An everyday word counts half: one most headlines write in lower case that more than one in a hundred of the stories carry (and at least ten); it counts whole once the two share a name. Stories are counted once the headlines are grouped with every word whole. A headline joins the story it shares most with, and starts a new one when none matches. Then two stories of two or more headlines each, their first headlines within 48 hours of each other, become one when the words they share count three the same way (a word is shared when two or more headlines of one story, or one in five of a big story's, and one of the other's carry it); they share a name, or a pair of words side by side with a name in it that two headlines of each carry and no more stories than a name does; two of the shared words or pairs are as rare and are carried by enough headlines of both stories, or one when four or more words are shared; and the earlier first headline carries two of the shared words, one of them rare. The bigger story takes the smaller. A story's heading is its first headline, in that outlet's own words. Outlets are counted once each; a headline repeated word for word by several outlets is counted as one headline.
- How lines are ordered
- In the feed, the newest date first; on one date, the story with more lines that day first (a count Browse makes, printed as DERIVED); then Wikipedia's own order. The brief's deck is ranked by the rule it prints beside it: The deck leads with the stories the most sources carried in the last day. Each registered publisher whose own feed has a headline on the story in the last 24 hours counts one; Wikipedia's editors count one when they added a line under it on today's or yesterday's Current events page, and one more when they posted it to In the news in the last 24 hours. A publishers' story joins a Wikipedia story when its first headline shares two or more long words with the story's name, a name of several words counting once. Ties go to the story that moved most recently, then to the one with more lines. Two stories on one event show once: a story whose name shares two or more long words with a story above it (by the same rule, a year not counting, nor a word one name writes as a name and the other as a common word) folds into that story, which says how many more outlets the folded one adds. A shared picture alone does not fold a story, since one file photo can picture several events. A story no source moved in that time leaves the deck, unless fewer than three moved. The counts are Browse's (DERIVED). A publishers' story with no Wikipedia match is a deck story of its own, headed by its earliest outlet's own headline; Browse writes no title for it. Until 22 October 2026 In the news's own order led the deck, and its top item could be days old. Just in, above the deck, shows the newest stories two or more of the registered publishers' own feeds carry, from the last six hours, each headed by its earliest outlet's own headline. No story is promoted by Browse's judgement, and no count is a ranking of importance.
- What Browse works out
- Counts (lines, citations, outlets, owners; the pages and sites in GDELT's newest file), each cited outlet's class (an official record, a party's own site or account, a social post, or a report, from the cited address and Wikidata), who runs an outlet as Wikidata states it, when Wikipedia's editors first listed a line and how its words changed (their revision history), and, on request, whether a line's quoted words are on the cited page and what that page declares about access and its own time. Every one is labelled DERIVED or named as the source's claim, with its date.
- A story's card
- A story the outlets' feeds carry opens a card of its own (/news/c/...). Its heading is one outlet's own headline: the one whose words the most other outlets' headlines share, ties going to the earlier. Browse takes out page furniture by printed rules (a LIVE or BREAKING tag, emoji, the outlet's own name on the end, capitals, spaced dashes, a "here's what to know" tail), changes no other word, and says under the heading whose headline it is and what it took out. What the feeds show is sentences filled in from the same feeds, labelled DERIVED. No model writes any of it and no article is read.
- What can go wrong
- Wikipedia's editors can file a line under the wrong heading, edit it after Browse read it, or remove it
- A cited page can change, go behind a paywall or vanish
- An outlet's share picture is what that outlet chose for its article, which can be an older photo, and the printed rule can put it beside another outlet's headline or a Wikipedia story it was wrongly joined to, which is why the credit names whose picture it is and links its article
- The heading on a story's card is one outlet's headline, and when the rule groups related reports (several races on one election night) it names only one of them
- Wikidata's owner or country for an outlet can be stale or missing, and an address can resolve to the wrong item
- A day page with more than 250 revisions shows first-listed times as "by", not "at"
- A party's own site can be matched to the wrong party when two share an address
- A newsroom's article can carry wire copy (Voice of America's AP, AFP and Reuters material is not public domain), which is why nothing of any newsroom's text is shown
- A page's own time, as GDELT recorded it, can be wrong by its publisher's clock, so one more than two days before GDELT's file is set aside and one later than Browse's read is not used for order
- A title GDELT recorded with a broken character is left out and counted
- GDELT's title for a page can be the site's name rather than the article's, and a place GDELT read can be the wrong one
- GDELT's crawl reaches few of the big outlets (0.9% of its records came from the twenty studied)
- An issuer's feed can answer with nothing for a moment (the Federal Register's did on 12 October 2026), so a feed that fails or answers empty keeps its last good read
- A mute by an outlet's cited name can miss a line that cites the same outlet by another name.
- Browse prints what it could not read rather than a zero.
- Corrections
- A fact in a line belongs to Wikipedia's editors and the outlet they cite; correct it there. A problem with how Browse shows it: every story page has a report link that names the page, read at comboventures+browse@gmail.com. What changed on a story page is Wikipedia's public log; what Browse changes in how it reads is in its ARCHITECTURE record.
How Facts are picked and checked
Facts is the one part of Browse written in its own words (decision 0016). A fact starts as a candidate: a hook from Wikipedia's "Did you know" archive, or something surprising found while reading elsewhere. Being repeated on several sites does not make it true, so each is checked against the article and the source that article cites, and dropped when they disagree or the source cannot be read. Only then is it rewritten, keeping every condition, date and unit, and the new wording is checked against the source again. A fact is also dropped when it is no longer true, too well known to surprise anyone, too broad to check, or promotional.
Each fact says whose words it adapts ("Adapted from Wikipedia", with the article, the CC BY-SA 4.0 licence and a note that it was reworded, or "Written by Browse"), lists its sources, and says who checked it and how. A check by an AI research agent is labelled as one, never as a person's review. When Browse's editor has read a fact and chosen it, it says "Selected by Browse's editor"; only after the editor has inspected the evidence itself does it say "Sources checked by Browse's editor". A test also flags any figure in a fact that none of its quoted passages holds: a warning, not a check of the claim, since a number can appear in a source and mean something else. Every fact has a permanent address and a way to report an error. During the pilot the fact pages are kept out of search indexes.
What Browse never does
- Browse writes none of the news: every line is its source's, with the outlets its editors cite.
- Browse never invents a figure: a number is a source's, for the source's time, with its kind beside it; what Browse works out itself is labelled DERIVED.
- Browse has no accounts, no cookies, no analytics, no advertising and no third-party scripts: what you follow and save stays in your browser (/privacy). Three other sites' players open inside a page only when a reader taps one: Windy's webcam timelapses, NASA's Eyes apps, and YouTube's player for the home page's chosen videos. In News, a story can show an outlet's own picture, which the reader's browser loads directly from that outlet with no referrer; a switch on the page turns them off.
The pages
- Now: The world right now: a globe with the day's earthquakes, alerts, storms and the space station; what changed since your last visit; the day's discoveries worth browsing; ways in to views, history, world stats, places and space; live webcams where it is daytime now; your places' local time and weather; what's happening now in events, news and attention; and ways to follow your curiosity.
- Discover: One thing worth knowing at a time: the strongest earthquake, the next launch, the nearest asteroid, today's picture, what happened on this day; each with why it was chosen, its time, its source and its kind.
- News: What is happening and who is reporting it: Wikipedia's In the news and Current events, the publishers' own feeds grouped by a printed rule into stories (Coverage; who is read and how an outlet is removed at /news/publishers), the headlines of six newsrooms whose licences or terms allow reuse (Voice of America, Global Voices, The Conversation, ProPublica, UN News, Inside Climate News), the pages the GDELT Project's crawlers read in their newest quarter-hour in English and 65 other languages, and governments' own records (the White House, the Federal Register, the European Commission, the UK Government, the European Central Bank) with companies' current reports as the SEC's EDGAR lists them, each line with its source and each cited source marked by who published it (an official record, a party's own site or account, a social post, or a report). Every story has its own page at /news/story/<Wikipedia title>, with what changed (read from Wikipedia's edit history) and an Atom feed at /news/story/<Wikipedia title>/feed.xml.
- Places: Any place on Earth: its local time, the nearest observed weather, active alerts and recent earthquakes nearby.
- Money: Exchange rates against a base you choose, from the European Central Bank's daily reference rates, with their change and what 1 unit buys.
- Planes overhead: The aircraft over a place you choose, as adsb.lol's volunteer receivers report them (open data, ODbL 1.0): a scope centred on the place and a list, nearest first, each with its callsign, type, altitude, speed, distance and how old its position is. Read again every 30 seconds while the page is in view; no position is moved forward.
- Space: The Sun now and the Earth from a million miles (each with its newest picture), the International Space Station's position and visible passes, upcoming launches, the crew in orbit, and space weather from NOAA.
- The Sun: The Sun as NASA's Solar Dynamics Observatory photographed it minutes ago, in eleven channels with NASA's own words for each; a replay of the last 24 hours or of any day since 2016; the corona from NOAA's CCOR-1 coronagraph; and the flares, coronal mass ejections and geomagnetic storms NASA's DONKI records, each opening the moment it happened, a flare's location marked on the disc. Every picture with its own time, from its file name.
- The Earth from space: The whole sunlit Earth from a million miles away, as NASA's EPIC camera on NOAA's DSCOVR recorded it through a day, played in order, in natural or enhanced colour, any day since June 2015, each picture with EPIC's own time and its age; and yesterday's whole Earth from the NOAA-20 satellite's passes.
- Space in 3D: Interactive: the Sun turned all the way round with NASA SDO's newest picture on the side the Earth sees and DONKI's flares of the last three days where they happened (the far side left plain); the eight planets at any moment from 1800 to 2050, computed by Browse from JPL's orbital elements, with time to play and distances from the Sun and the Earth; eight of NASA's 3D models (the space station, Hubble, Parker Solar Probe, Chandra, Ingenuity, the Apollo Lunar Module, the Saturn V, asteroid Bennu) in Browse's own viewer; and NASA's own Eyes apps, loaded inside the page only when a reader taps.
- On this day: Today through history: the newspaper front pages of a hundred years ago, who died this week, the Nobel Prizes as they are announced, new films and games, and everything that happened, everyone born and everyone who died on the date, the best-known first. Any other date is /history/MM-DD. Every front page opens in Browse's own newspaper viewer (/history/paper/...): zoom to the full scan, a magnifying glass, find any word, and highlights and notes kept on the device.
- Died recently: Who died yesterday and in the days before it, from Wikipedia's Deaths in month lists: each name with the age and the words the list gives, and the report its editors cite (CC BY-SA 4.0). The best-known first within a day, by the Wikipedia languages with an article.
- TV tonight: What is on TV in a country you choose, and what is new on streaming services today, from TVmaze's published schedules (CC BY-SA): each episode with its channel or service, time and number, a new series or season marked. Times are SCHEDULED, in your own time zone.
- Sport tables: League tables for the Premier League, La Liga, the Bundesliga, Serie A, Ligue 1, the NFL and the NBA as Wikipedia's editors record them, with their updated-to date, their source and the revision read (CC BY-SA 4.0). No live scores, no odds.
- Compare: Countries side by side on the World Bank's published figures: life expectancy, money, population and more.
- World stats, live: The world in numbers as it moves: births, deaths and the population as estimated counters with the year so far and the rates, the Global pulse, earthquakes and alerts counted, exchange rates, the US debt and world population, space weather, the space station's position, the people in space and the next launch, attention on Wikipedia, and what has ended.
- The live map: Earthquakes, alerts and storms where their sources put them on one live map, with the news beside them, the world's pulse (four key numbers) and what changed since your last visit; filters, time windows and a replay.
- Live views: Public cameras and webcams, as their owners publish them: near your place, the world's most-watched webcams by kind of place (Windy.com), Kīlauea's summit (USGS) and California's highways (Caltrans), each still with the time it was made. Public places only; nothing recorded.
- Live wall: A full-screen wall of live counters, feeds and camera views for a screen in a room; /live?stream=1 fits it to a streaming app's browser source.
- News feeds: News as Atom feeds: the front (/news/feed.xml), each topic (/news/topic/<key>/feed.xml) and any story (/news/story/<title>/feed.xml); every entry one of Wikipedia's lines with its attribution, the outlets its editors cite and when they first listed it.
- News publishers: The publishers whose own feeds the Coverage view reads: what is read of each (link, headline, time, author, categories, the picture's address; never the text), what each outlet's own terms say about its feeds, when it was measured, the weekly check of its rules for pictures, and how an outlet is removed, or its pictures or small icon taken off, within 48 hours. Each outlet's current headlines are at /news/outlet/<key>.
- Saved: What a reader saved, followed or marked, kept in their own browser; Browse receives none of it.
- Privacy: What Browse keeps (nothing about you on its servers) and what stays on your device.
- Terms: The terms of use.
For machines
Agents and crawlers are welcome: a short map at /llms.txt, everything in one file at /llms-full.txt, what an agent may do at /agents.md, and the pages at /sitemap.xml. Search engines may index Browse's main pages; a page too thin to stand alone asks them not to.
Last updated 5 Nov 2026