Authentication & limits¶
Let's get you connected! The first thing you'll want to know is how to authenticate your requests. While you're here, we also encourage you to at least skim the documentation on limits.
You can get started developing locally without any special effort or cost. If your application
is running via a development web server accessed via
127.0.0.1, we'll treat your
requests as coming from a free plan. Note that this is designed for convenience of development on
a single machine, and is rate limited accordingly. If you start receiving HTTP 429 responses,
please sign up for an account (it's free!) and use an API key.
Adding a domain to your property in the client dashboard is the easiest form of authentication for production web apps. No additional application code is required, and you don't need to worry about anyone scraping your API keys. We recommend this for most browser-based applications.
You will need to include an API key with each request that is made outside a web browser. You can generate a key in the client dashboard. Please take care not to expose your API key. You generally should use domain-based authentication for web browser applications. Use API keys only for requests coming from outside a web browser or for local development when you need to make requests in excess of the rate limit for keyless access.
Note that security of API keys is your responsibility, and requires special care whenever the application has end users outside your organization. While major mobile platforms like iOS encrypt and sandbox applications, all bets are off on a jailbroken device. You can mitigate these risks by having your app pull an API key from your servers at regular intervals, ideally via an authenticated endpoint. Always store secrets securely whenever possible. For mobile applications, most platforms have a secure credential storage and retrieval API, and we highly recommend using that.
You can use API keys in two ways. The simplest is to add a query string parameter
api_key=<key> to your request URL. For better protection of your API key (from certain logs at least),
you can add an
Authorization header with the value
Stadia-Auth <key>. In both cases,
is your API key (without angle brackets). You can generate and revoke API keys in the
Authorization: Stadia-Auth YOUR-API-KEY-GOES-HERE
We have two forms of limits: the features you can access with your plan and request-level limits. You can see the list of features included with each plan on our pricing page. In this documentation, we try to highlight when a feature is only available on certain plans with something like this (which indicates that a feature is available only on paid plans):
Our convention is to not include a plan grid if a feature is available to all users. If the feature(s) discussed on a page are only available to a subset of plans, we will either put a plan grid at the top of the page (if it applies to the entire page) or under specific headings where limitations apply.
Our plans each come with request limits as specified on our pricing page. After your request credits and grace threshold are exhausted, our default action is to hard limit until the start of the next billing cycle or plan upgrade. This protects you from unexpected bills. If you want to keep serving requests beyond the plan limits, we offer opt-in pay-as-you-go billing with simple, discounted pricing on the additional credits. See our pricing page for details.
We will send an automated email if your account has been hard limited.
Free tier restrictions¶
We offer a free tier for development, evaluation, and non-commercial use. Other uses of the free tier are not permitted and require a paid subscription.
We reserve the right to impose rate limits on customers who are abusive or disruptive of service. We also enforce rate limits on local development as noted above.
We impose the following limits universally across on our Geospatial APIs. If you have a more demanding use case, send us a message, and we can discuss your use case.
|Costing/Mode of Travel||Max Route Locations|
|Automobile and truck||20|
All matrix requests have a limit of 1000 elements, regardless of the costing/mode of travel. A matrix request with 3 inputs and 5 outputs has 3 x 5 = 15 cells. This means you could send a 10 x 10 or 2 x 50 matrix request (each having 100 cells), but not 40 x 30 as it has 1200 cells.
All matrix endpoints (including optimized routing) have a limit of 1000 elements, regardless of the costing/mode of travel. A matrix request with 3 inputs and 5 outputs has 3 x 5 = 15 elements. This means you could send a 100 x 10 or 20 x 50 matrix request (each having 1000 elements), but not 40 x 30 as it has 1200 elements.