Yandex.Weather

2025: Ability to chat with a neural network

Yandex Weather has opened the opportunity to chat with a neural network. Now the service not only informs users about the weather, but also helps them make decisions taking into account weather conditions. The chat is powered by Alice AI technology and can seek advice - such as how best to dress before going outside. The chat will also help the user make decisions about how to spend the New Year or vacation depending on the weather. For December 2025, chat is available to a small percentage of users in test mode. Yandex announced this on December 29, 2025.

When the user asks a question in the chat, Alice AI technology analyzes more than 100 meteorological parameters and their combinations: from basic (wind, temperature) to complex (allergens). To do this, the neural network was connected to the Yandex Weather server using a special MCP (Model Context Protocol) protocol. This allows neural networks to receive up-to-date data on changes in meteorological conditions. The neural network has access to data that is not visible in the service for users, for example, an archive of weather data, which allows you to find out the weather in previous days.

The neural network gives the user a detailed response based on the current weather conditions in his city or district. It analyzes the current and upcoming weather and gives suitable advice. Chat allows you to answer difficult questions. which would require a person to carefully and for a long time analyze the data presented in the product.

2024

OmniCast Technology Launch

Yandex OmniCast technology has been launched in Weather. Neuronets improved the local forecast through new sources. data Yandex announced this on August 23, 2024.

OmniCast takes into account data from both professional and amateur weather stations. Thanks to this, the service has become 36 times more likely to receive temperature data and give a more accurate forecast for each city quarter.

OmniCast is part of the Meteum technology that underlies Yandex Weather. OmniCast uses not only weather forecasts from global suppliers, data from radars, satellites and ground weather stations, but also information from amateur weather stations. They are located everywhere and transmit information every five minutes, while professional ones - every three hours. This made it possible to receive data more often and recalculate the forecast. OmniCast works on the basis of neural networks that calculate data from nearby weather stations, temperature fluctuations in the microdistrict and its features. As a result, users receive a forecast for each plot of 500 by 500 meters.

The local forecast is especially relevant for large cities, where each district has its own microclimate. For example, in the city center it can be 13 degrees warmer than on the outskirts. According to OmniCast, in the summer of 2024, the hottest areas of Moscow were Arbat, Zamoskvorechye and Tverskaya district, and the coolest inside the Moscow Ring Road were Bibirevo, Konkovo and Bogorodsky.

Ability to generate weather dynamics

Neural networks in the updated version of Yandex Weather for business will predict the dynamics of weather indicators. This will help companies of all sizes plan their activities more efficiently. Information is available to enterprises in their personal account in the form of an API or weather dashboard, which allows you to control weather data for more than 150 parameters and several locations at the same time. Yandex announced this on March 5, 2024.

The business will also be able to get access to the service to the seasonal forecast for six months, API with weather data for the next 30 days, hourly historical weather data in Russia since 1950 and an individual industry solution from analysts and meteorologists. Yandex Weather for Business uses its own Meteum technology. It combines classic meteorological approaches with algorithmic and machine learning, analyzes data from five independent meteorological models and takes into account indicators from radars, satellites, ground stations and user messages.

𝓲

Фото: Yandex

Weather information anywhere in the country with an interval of one hour will be useful for analyzing more than a hundred different business scenarios, including those created independently. You can integrate the tool into your own analytical or CRM system. For retailers, the data will help track, for example, how long precipitation increases demand for delivery. And energy companies using the tool will be able to adjust the forecast for electricity demand depending on the average monthly temperature of the winter season.

Шаблон:Quote 'author=said Alexander Ganshin, head of Yandex Weather.

2023: Data integration with Yandex Routing service

Yandex has learned to take into account the weather when calculating routes for business. This became possible thanks to the integration of Yandex Weather data and APIAndex Maps: now the travel time will be adjusted taking into account historical data on the impact of weather on the speed of cars. At the same time, the actual weather forecast and the situation on the roads will help to clarify routes in real time. Yandex announced this on November 1, 2023. Read more here.

2022: Launch of thematic forecasts

On April 28, 2022, Yandex reported that Yandex Weather began to give thematic forecasts to those users whose hobbies depend on weather conditions. The first was a detailed weather forecast for gardeners. It is available in the Yandex application and is calculated using Meteum technology.

Thematic forecasts will contain information important for a particular occupation. For example, gardeners will be able to find out what the moisture and soil temperature are and whether there will be more frosts - this will help plan plantings and decide whether it is time to open greenhouses. Users do not have to search for this information in different sources. Everything you need can be found in one place - in the Yandex application.

"We have released a forecast for gardeners for the May holidays, because many will go to the dacha. In the future, there will be more such forecasts. Data for water sports enthusiasts is already being tested, followed by a forecast for fishermen, tourists, cyclists and runners. Thanks to Meteum technology, they will be able to learn about the weather anywhere on the map, "- told the head of the service Alexander Ganshin.

Meteum technology is based on machine learning and calculates the weather forecast with accuracy to the microdistrict. The technology uses data from not only meteorological instruments and weather stations, but also millions of users of the service.

See also

• Climate in Russia