Tomorrow.io

Mexico City, Manila, Boston — three runners, three completely different skies this morning. All three deserve to know whether their run this morning is going to be confronted with a thunderstorm, or if they’re running in a plume of wildfire smoke and just hadn’t checked the news. Zeev Blumentzvaig, one of our engineering leaders with a Garmin watch, was motivated enough by his watch this week to ship a Tomorrow.io app on the Garmin platform. It answers the two questions every runner asks before lacing up: → Is it going to rain in the next hour? → When's the best window to run today? Check it out here: https://lnkd.in/g96JU5sh Simple questions. What's hard is where you're standing when you ask them. Most wearables lean on traditional sources for weather, which can be dense in some cities, but thin or missing the moment you're on a coast, in the hills, or in the majority of the world's cities in which a network like this was never built to reach. Tomorrow.io reads the sky through our own constellation of satellites, everywhere. So the rain call in Manila this morning isn't a guess borrowed from miles away. To the Garmin team — this is a small taste of what opens up when global weather meets the outdoor enthusiast who needs information like this to make a decision. We'd love to show you the rest. 🌍 Runners: what's the one weather question you wish your watch could answer mid-run?

A few days ago we showed you a river in the sky from the Tomorrow Constellation. Today: its the opposite. This week a ribbon of hot, bone-dry Saharan air rode out over the tropical Atlantic - the Saharan Air Layer. You may have seen the dust in photos: hazy skies, red sunsets. But the dust is the passenger. The driver is what you can't see - a slab of desert air roughly 4 km deep with the water wrung out of it, starving tropical waves of the moisture they need to become storms. That invisible part is what our sounders measure. Cameras photograph the dust. A microwave sounder reads the water - and its absence. Over four days, our sounders have made +180 passes over this stretch of ocean - about 2 million soundings. The dry mass marched roughly 1,000 km west, grew six-fold to over 1 million km², and pulled column water vapour down to ~21 mm in a basin that runs mid-30s. Context that matters: by mid-July the Atlantic has usually named its second storm. This year it's named one. Forecasters put a strengthening El Niño first in that quiet - with this dry air as the supporting act.

It's easy to underestimate how many jobs weather intelligence is actually involved in. Predicting whether a flight will divert. Calling the moment it's safe to send a crew back outside. Rescheduling a fleet to beat a storm. Protecting a construction schedule from the one afternoon that breaks the concrete pour. A few examples of thousands. Each carries its own problem, its own decision, its own people trying to navigate the uncertainty — and its own way of conveying the situation and delivering a solution focused on insight and intelligence, not just a weather forecast. A decade of solving for this is what has made the Tomorrow.io platform remarkably flexible and capable. Because it had to be. Labs is where that flexibility keeps compounding: more refined, more focused experiences, tailor-made to the individual use cases that drive ROI for our customers and impact for the world. More custom solutions as their needs evolve. And more nimbleness — for our team and theirs — to react and solve problems like never before. The distance between "we need this" and "we have this" keeps getting shorter. https://lnkd.in/gz9iMjNW

In April we shipped the Tomorrow.io MCP server — weather as a set of tools any AI agent can call. Last week, Gale — the AI agent inside the Tomorrow.io platform — became now available via the Tomorrow.io API. It's the same agent that reasons about weather for our own users. Now you can embed it in your product: hand it the question — "will the storm reach the venue before gates open?" — and it answers in your user's language, units, and timezone. No translating the question into technical parameters first. Gale is the newest of three ways to bring weather into what you're building. The clean way to tell them apart is who does the thinking — and what lands in your app: Gale API — our agent, in your product. You ask, it reasons, you get an answer. Delegate the weather thinking to your platform setup. MCP server — your agent, our data. An MCP layer on top of the Tomorrow.io API: the fastest way for your agents to pull a forecast as a native tool call. You keep the reasoning; we make the data trivial to reach. Custom Vector Events — your app, our payload. Weather arrives as a rich vector object — event footprints, geometries, context that carries forward — and your agents work it natively: spatial joins, intersections against your own assets, the analysis agents are genuinely good at equipped with a payload of context. One question — what is the weather about to do to my operation? — answered wherever it fits best: hand it off, wire it in, or compute on it yourself. The MCP's first three months show the range of use-cases: dispatch agents checking the next 12 hours before committing trucks, ops assistants folding site forecasts into morning briefings, no-code teams wiring thresholds to alerts through the n8n, Zapier, and Make connectors we've added since launch. Anyone can try it in ChatGPT now — we're on the OpenAI App Store, no API key required. https://lnkd.in/da_KMdHF

A river in the sky came ashore over Chile this past week — a plume of water vapor thousands of kilometers long, crossing the Southeast Pacific and breaking against the Andes. Forecasters called it an atmospheric river; it gave Santiago its wettest July in 20 years. We watched it from orbit — five days, across the ocean to landfall, with our own constellation of microwave sounders. The physics becomes visible side by side. In the river's core, most of the moisture rode in the lowest ~3.5 km. The mountains behind Santiago rise past 4,700 m — with peaks well above 6,000 m. The range does the rest: the same instrument that measured the vapor over the ocean watched the rain come down on the coast and the Andes, heaviest over the mountains. One storm, watched from approach through landfall. #AtmosphericRivers · #Chile · #EarthObservation

104 matches. 16 cities. Three countries. Put every game on one thermometer — next to the last five World Cups — and the story isn't the one you'd expect. → Warmest of the five, on average. But not the most relentless: 2014 in Brazil was more uniformly hot. 2026's signature was range. As the first sea-to-altitude, three-country World Cup, it swung from cool Vancouver nights to Houston afternoons. → The roofs did the real work. Houston read 104°F outside at kickoff — roof shut, AC on, the pitch felt closer to 72°F. The hottest match played under open sky was Philadelphia: Paraguay v France, 101°F. → For scale: the coldest game of the last five World Cups was a 2010 night in Johannesburg that felt like 26°F. This tournament nearly covered that whole span on its own. Every number is measured from the historical weather archive — for each match window, at each stadium. Full interactive breakdown, all 104 games hour by hour: https://lnkd.in/gmufyGyT

Many major leaps in AI this last decade came from the same lever: more data made dramatically better models. Weather never got that same scale-up. Not for lack of value, for lack of supply. Earlier this month, the eye of Super Typhoon Bavi passed directly over Rota - a Northern Marianas island of fewer than 2,000 people - with sustained winds near 180 mph. For all 8.5 days of Bavi's life - a 25-knot system to a 156-knot peak - the Tomorrow Constellation kept it in frame alongside the six US and European MW sounders. Across all 13 named tropical cyclones on Earth this year, Tomorrow.io's satellites took 1,713 microwave looks together with those six sounders' 732, additive, on top of every pass they flew. And 232 times this season, the public picture went three or more hours without observing the storm. 1,432 hours of stale sky in all. Tomorrow.io's sounders placed 1,250 looks inside those windows. All 232 of them - ensuring the storm never went unwatched.

The Open Championship is the only major that has always been shaped by wind. Links golf began on treeless coastline because that was the only ground nobody else wanted — sand, gorse, and weather off the sea. The wind wasn't added as a hazard. It was simply never blocked by the vegetation. 154 Opens later, Royal Birkdale still gets read the same way: hole by hole. So we read it and built The Wind Card to help you see where the advantage lies. Every hole's real tee to green bearing against the hour-by-hour forecast, scored with the wind penalty we measured across 263,553 tour rounds (+1.25 strokes per 10 mph, field-level). What Thursday looks like: → The wind costs the field about +4.3 strokes across the championship. → Holes 4, 5, 10 and 11 turn into the teeth through the 1–4p peak. → One gift all week: 17, a par 5 that turns downwind Friday late afternoon.

It’s a beautiful day. Few words are better. But what actually makes a day beautiful? Over the coming weeks, we’ll reveal: The cities with the most picture-perfect days each year. Where exceptional weather happens most often around the world. What days of the year are consistently the best? And a few surprising insights and experiences we think will change how people think about weather.

The sky has its own rush hour. It's 5 p.m. Experience the data at https://lnkd.in/gRUKrhM9