Show HN: Stadia Maps, reasonably priced map tiles, static maps, and routing
stadiamaps.com
stadiamaps.com
A client was struggling with Google Maps limits and pricing, and we saw a need for simple map services with even simpler pricing. With the recent announcements from Google Maps and others, we see that this is even more important than we initially realized.
We believe maps should serve the customer, not the map provider. Some mapping providers display maps with competitors to your business, reviews (that may be fraudulent), or their own logos alongside your data. We believe a well-designed base map is all you need to allow you data to shine. As we expand, we will continue to focus on putting your company and data first.
That's just the beginning; we are working on several simple tools to help business owners and developers utilize maps in their web and mobile applications. Ultimately, we want to enable others to create fantastic mapping experiences.
How
- We strive to provide fairly priced and transparent plans.
- We provide beautiful themes for a variety of contexts.
- We never track[0] or sell user data.
- We never display reviews or potential competitors (“related” businesses) alongside your data.
- We never display ads on your map.
- We are working on JS libraries to make interactive map design suck less.
[0] Beyond the necessary details to implement rate limiting and similar technical measures to ensure everyone has a good experience
As I was typing this, I visited your site again and realized that there's more to the website when I scroll down, but all the was lost below the fold. Clearly, a lot of work has gone into making Stadia and it deserves to succeed. May I suggest that you take another look at how users interact with your site?
Thank you for your kind words. We're definitely excited about the opportunities before us.
And thanks for the feedback!
I agree with your critique. We definitely need a better marketing website that conveys our mission and vision, as well as product offerings. I realized as I was reading my parent comment that most of the big selling points (which are things we strongly believe in) are not properly addressed in the website.
I think I know our next action item. :)
If you're just using tiles, I don't see any difference with Stadia Maps — perhaps price, I don't use OpenMapTiles' commercial services.
Stadia do offer routing using Valhalla, which OpenMapTiles don't offer.
(I have made some small contributions to the open source OpenMapTiles project [result at [3], vector/raster tiles in other projections], but have no connection to the company.)
[1] https://openmaptiles.org/ and [2] https://openmaptiles.com/
We're also trying to offer value in other areas such as the design, and in additional services. Routing is obviously one of these, and we've got a few more services that that we're working on. We'll be announcing some of these later this year.
I use Galileo Offline Maps on my iPhone, and my own localhost browser-based map viewer on Mac that can read SQLite tile stores and let me put address queries on the map, and draw circles around a point.
The tiles I use were cached once with MOBAC, when I had a fast Internet connection. (when I need a map, I usually don't have Internet access). Google Maps keeps redesigning and making roads harder to see, emphasising Places instead. Maybe that's good for advertising, but harder for me as a cyclist.
Stadia maps are using OSM tiles, but with a beautiful format, and pinyin romanisation for the street names. It looks just like Google Maps did back in 2014, when the colours were bright and clear.
I'd like to generate a large tilestore (usually ~2 GB for a whole country [1]) of the Stadia formatted OSM data, using MOBAC. Could someone suggest how to do that?
We'd love to help anyway possible with the generation. Would you mind sending us an email (info@stadiamaps.com) so we could collaborate better on this?
Then you could script calls to a renderer to convert.
• The height of the #use-case element can vary (for many mobile-sized displays, “locate nearby caffeine” takes two lines, unlike the rest which are one line), which messes up the scroll position regularly when you’re below it. A solution is to take the longest, have it there for layout with `visibility: hidden`, and have the actual text at the same point, but with `position: absolute` so that it doesn’t contribute to layout.
• .cc-window (the stupid cookie banner) has z-index: 9999 but #action has z-index 99999 which is higher, but should probably be lower. (Else it blocks the “Got it” button.)
• That map sure took a long time to load. Fifteen, thirty seconds for initial render (high-DPI display, Firefox, Australia). Apparent indefinite hang for some tiles when I was interacting with it so that I just gave up on it after a minute or two. Sometimes it worked better after a reload.
We'll look into those display issues right away. We pushed a new design before the HN post took off, and we clearly missed a few details.
Regarding map load time, our servers are currently located in the US (which is where all of our paying customers are based at the moment). We do have multiple servers running in a cluster, but we are still working on our geographic diversity. While our load times are generally good wherever we've tested in the northwestern hemisphere, my cofounder lives in Korea now and he is acutely aware of our latency spikes in Asia. Our CDN, CloudFlare, routes traffic based primarily on peering arrangements until you start paying them a lot more, but we're working on geographically diversifying our server fleet behind CF load balancers as our next step.
One thing I noticed (and imagine was an oversight):
Credit not given to the Vahalla Project.
The OSS Routing Engine "Valhalla" [0] is listed as the core of the Routing feature [1], yet doesn't seem to have made it onto the attributions page [2]. This not following (i.e. violating) the terms of the Valhalla license.
[0] https://github.com/valhalla/valhalla/blob/master/COPYING
I'm asking because I am a a longtime MapBox subscriber, they develop various client libraries, provide support, and there is interface for map styling. This looks like direct competition.
We believe there are advantages, though some apply mainly for specific applications:
- Price: We use tier-based plans and soft-limiting instead of purely "you pay for what you use", this allows for consistent bills, even if your traffic isn't always consistent. Thus it's very easy to see how much it would cost each month, without a fancy calculator and ready knowledge of how users behave.
- Asset Tracking: We do not require $500 / month plus usage to use our maps in an internal or asset-tracking application.
- Goals: We want to build a sustainable, bootstrapped business without the requirements of venture capital, etc. This is a key philosophical difference. MapBox is venture-funded and recently raised again, while we're bootstrapped and don't plan to raise VC money. This allows us to exist by providing a quality service to our customers without the need to grow exponentially or incessantly. (This is not a knock against venture funded companies; we just believe there's more than enough room for both philosophies in map services companies.)
This could be really important. Mapbox is a Softbank rocketship, which is trying to get to the stratosphere but has a big chance of exploding on the way up. If they can't figure out how to become a $10B+ public company, best case they are bought by e.g. Google and shut down 14 months later.
Whereas if Stadia can hit break-even and become sustainable, they could have a smaller but more reliable platform that sticks around.
https://docs.stadiamaps.com/vector/
Chrome Version 66.0.3359.139 (Official Build) (64-bit) Window 10
So, the short answer is, yes, you could do this with OSM's data, but most companies prefer to pay to abstract away the management efforts.
We also provide other geo services, such as static maps and routing, which are often useful for companies and projects needing maps.
Sure. Here's a brief overview:
- Bill-as-you-go vs tier-based billing Many other services use usage-based billing, which is great for minimizing costs, but not so much for cost-prediction. We use tiers with soft-limits, so if you exceed your given usage for a few days of the month, it's not big deal and your bill doesn't change (e.g., you hit the HN front page!). Our initial customers told us consistent costs were very useful when handling client projects.
- Raw price We believed we could deliver a premium product at lower costs than other providers. Based on price comparisons, we believe we are on the right path to achieve this. We also do not treat certain applications (such as asset tracking) differently.
- Map design We didn't like many of the other provider's themes, so we built our own and plan to expand our portfolio over time.
- Goals We want to build a sustainable map service provider that respects the rights and data of its users (and its users' users)—not grow at any cost. Many of the other providers aggressively market for growth or capturing users' data, or marketing their own products on the map (e.g., reviews or other places), we want to get out of the way and allow your data to shine on our maps.
There are more differentiators, but these are some of the factors that led us to begin our own service.
I have an idea in the works and was beginning to research self-serving OSM tiles for it. Will definitely try Stadia.
Yes and no. The OSM project provides a database dump. You can download it all ( https://planet.openstreetmap.org/ ). But it's not really in a format that you can draw maps from without processing (look at https://switch2osm.org/ to see how).
> OpenStreetMap doesn't charge to use their service commercially iirc
The map images on openstreetmap.org are run by unpaid volunteers, in a donated server space, and is for the OSM project and mappers to use. If some group starts hammering the service, then they'll block you, commerical or non-commerical is not relevant. There is no guarantees of service. If you want guarantees, pay someone :)
Case in point: have several small admin applications, Mapbox and Google want $500+ for them because they're internal and restricted access.
We do require a paid tier because almost all asset tracking applications are commercial, which based on our non-commercial/development-only clause for the Free tier requiring a paid plan is the logical outcome. If you have a non-profit project that does asset tracking and is under our Free limits, please let us know, and we can discuss options!
That is definitely a bug in our pricing table. We'll correct that.
As an aside, are you are of any method for distributing changes to the OSM database as they occur? Failing a live stream, something like an hourly torrent?
It seems to be simplifying coast/river edges into straight line segments a bit too much? It's subtle, but I'm comparing to the default openstreetmap.org renderer.
We'll look into ways to remove the terms issue while preserving out ability to properly bill and protecting privacy and regulatory compliance.
Would a solution that allows you to request via your own servers (simple proxying?) meet the requirements? Or would a simpler solution be to allow you to generate (perhaps via an API?) and self-host, while paying for a service that allow this?
India, Pakistan, Ukraine, Russia etc needs to be handled differently to avoid users hate mail.
Could you elaborate a bit more on what you mean? Do you mean rendering the map differently based on a user's likely view of disputed borders, etc.? Basically smoothing over various geopolitical issues by making the map agree with a given user. :)
'Wrong borders' has given some governments reason to block access to sites.
Very difficult problem to solve unfortunately.
If you don't mind elaborating, we're curious what you are looking for that the "platform-native" maps from Apple and Google don't provide (both Google and Apple provide unlimited unlimited free usage on mobile). I (one of the founders) am actually a mobile developer for most of the day, and so far I only have one client that isn't using the built-in maps (Mapbox in their case). They wanted more recent high res satellite imagery and the ability to download areas for offline use. Are there any other use cases we're missing?
Would you prefer this be provided as a downloadable dataset or a linkable dataset to use just like our map themes?
I'm assuming its cloudy/cdn backed, so it becomes both local and cached.
If you're using vector maps, you can definitely remove that layer, if you wish.
If enough users request a basemap without country labels (we haven't had any requests for that before this), we certainly could add it for raster.
Thanks!