RainViewer has informed us that their radar API will stop Jan 2026. RainViewer has decided to focus on their app instead of the API - which is understandable. App development and web services are damn hard.
RainViewer has supplied Flowx with radar reflectivity data for many, many years. A big thank you to Oleksii and the RainViewer team for their data, service and help.
If you are a silver or gold subscriber, please be aware of this change in terms of continuing your subscription.
Does Flowx plan to replace the radar data?
At this stage, we have no plans. My current focus is on short-to-medium term forecasts and data at altitudes.
Just as a data point, Iâve used Flowx for years, and I was surprised to learn that it has real-time reflectivity data. I turned it on, and to my eyes it gets in the way of the forecast images. I expect most Flowx users use more than one weather app, and that most of us use other sources for real-time information.
Not replacing RainViewer in order to focus on forecasts, your central feature, is a decision that seems reasonable to me.
Of course, this is not to criticize sisters and brothers in the Flowx community who use this amazing app in different ways than I do.
Yes. The job of sourcing this data from many different sites in different formats at different time intervals, converting them to a common format, cleaning them, stitching them together, tiling them and then serving them is not a simple job. And then managing this on top of sites falling over or changing their format is not simple.
I already do this for weather data and itâs not simple.
I understand, Duane. Thanks for clarifying. Iâm puzzled by the number of weather apps covering the globe. Are their developers all doing the composition of a real time consistent global weather data set themselves?
Most use another service. Windy started by using RainViewer data. I think they process their own now or pay for another service.
After I replied your last post, it dawned on me that weâre talking about creating another RainViewer. Note, RainViewer has many developers. Flowx has one.
This is exactly the reason I donât have any plans to replace it yet. There is much we can do that fits nicely in the domain of weather simulation data.
Be great at one thing, rather than being mediocre at many things.
I use radar data mainly for a quick (visual) validation of Flowx results. It helps me to select the best data source at the time of interest. Here in The Netherlands I usually choose ICON-D2, sometimes HARMONIE. (While travelling, other sources are used of course.) Apart from radar data inside Flowx, I sometimes use Rain Radar for this quick check, as it provides some forecast data as well . Just a question, assuming Flowx is able to use the RainViewer API: if a user has RainViewer (and Flowx) on his phone, could a coupling with Flowx easily be established?
Yep, sad for the Flowx users, but as Duane said - understandable.
I long had hopes of getting the RainViewer predicted radar into Flowx for people like me who pay premium for both services. That is clearly not going to happen.
NexRad? Your US subscribers would be very happy. For continuous radar data provision in Flowx, five months to go (until end of RV support) is an awfully short period for development (at your excellent quality level).
Appreciating the work aspects etc. I would like to highlight the importance of Radar/reflectivity for some users.
One limitation of some forecast output is a probability prediction. eg UK Met Office gives a forecast eg âlight rainâ with a probability eg â60%â. So I treat any forecast as a prediction. So I might check several forecasts as consistency suggests higher confidence in the predictions.
Some geographic areas have additional effects often not well forecast (eg UK East anglia rain will often reduce to zero as rain approaches).
Rainfall radar/reflectivity provides an excellent means to assess short term (eg this afternoon, maybe 4 hrs ahead) rainfall predictions. Often when eg UK Met Office eg moderate rain 80% starting in an hour check the rainfall radar/reflectivity and you can see itâs not going to happen as there is no rain nearby on approach - so you can go for cycle, paint front door or whatever very confident there is no rain despite forecasts.
You can totally use the RainViewer app (or the many other radar apps) for what you describe. It also comes with forward predictions of radar/reflectivity which we didnât have in Flowx.
Yep, agree with Duane.
It seems RainViewer will stop providing radar data to wx forecast apps as they have entered the medium term forecast area themselves. So far, RainViewerâs GUI and level of user control is no match to Flowxâs in this medium term forecast area, which AFAIK is Flowxâs prime domain of application. For nowcast, I can choose between a free (KNMI) or a cheap (Buienalarm) local short term forecast app, both of which provide real-time radar data as well.
Hi Duane - I will really miss the rain radar feature and will probably cancel my Gold subscription.
The ability to compare the forecasts with the actual movement of rainfronts is a powerful feature that helps me to judge which forecast model is the most reliable and which best works for me for local purposes - the Flowx user âswipeâ interface makes this so easy.
Iâll probably subscribe to Rain Viewer, or a similar source, but having to use two apps will be a pain.
Iâd love to see you spend some energy and time of finding a replacement, or renegotiating with the RainViewer folk, I for one would pay extra.
Pay for radar data. It already exists. <â I canât justify spending money on radar data.
Process radar data myself <â this is a really really dumb idea because it will kill Flowx.
Donât replace radar.
I have limited time and money so itâs important to prioritise. This is why options 1 and 2 donât make sense.
Flowx is a weather app that shows weather simulation data and it does this well. It should focus on its strengths.
There are many radar apps out there. Several with 10+ million users. Flowx has 500k users. Why should Flowx spend time and money implementing radar to replicate what these successful apps already do?
I should use my limited time adding data sources (e.g., the HAFS hurricane model) or data at altitudes. This would strengthen what Flowx is already good at.
I should spend the money I donât have buying weather simulation data like the UKMO model or the Australia BOM ACCESS models. This would strengthen what Flowx is already good at.
Also consider this. Radar data would likely cost thousands per year - letâs say $2000 per year (probably way more). This is equivalent to 100 subscriptions. Will Flowx lose 100 subscriptions due to radar disappearing? Probably not, because people donât pay $20/year for the radar feature in Flowx when there are bigger and better radar apps out there. In conclusion, Flowx will lose money.
UKMO data costs around $2500 per year. Will I get 100-200 subscriptions to cover this cost? I donât know but more likely than radar since people buy Flowx for this kind of data.
At this stage, it simply doesnât make sense to spend time or money, I donât have, on radar.
On one hand, as a developer myself, I totally get it. Spending time and money on a feature that isnât the core of your product isnât often the best idea. However, again as a fellow developer, I also know the importance of feedback. And I know Iâm just a single datapoint, but what the heck.
I use Flowx for the UI/UX. Being able to have multiple layers with the same UI is the key feature for me. Thatâs the core. I can get forecasts (from actual trained meteorologists instead of armchair ones like me!) and data from other sources, but all the UIs for those sources doesnât click with me the way that Flowx does. In essence, I pay for a subscription because it makes my life easier; but that evaporates once I need to start using multiple apps.
Iâm not sure how many in your user-base are like me. Maybe Iâm the only one. I just wish one of the alternatives in the Rainviewer announcement were feasible as a replacement.
Radar data incorporated in Flowx is great for validation purposes and very short term forecast, but it is not part of its core function, which is displaying simulated wx data. Of course, like all users, I regret radar will no longer be part of Flowx as of next year, yet even then it will remain the preferred app for wx forecast, not nowcast, to me. Cheers.