An Introduction to APIs
Have you ever wondered how Facebook is able to automatically display
your Instagram photos? How about how Evernote syncs notes between
your computer and smartphone? If so, then it’s time to get excited!
In this course, we walk you through what it takes for companies to link
their systems together. We start oﬀ easy, deﬁning some of the tech
lingo you may have heard before, but didn’t fully understand. From
there, each lesson introduces something new, slowly building up to the
point where you are conﬁdent about what an API is and, for the brave,
could actually take a stab at using one.
Who Is This Book For?
If you are a non-technical person, you should feel right at home with
the lesson structure. For software developers, the ﬁrst lesson or two
may feel like a mandatory new employee orientation, but stick with it –
you’ll get your ﬁll of useful information, too.
Table of Contents
Authentication, Part 1
Authentication, Part 2
APIs (application programming interfaces) are a big part of the web. In
2013 there were over 10,000 APIs published by companies for open
consumption 1. That is quadruple the number available in 2010 2.
With so many companies investing in this new area of business,
possessing a working understanding of APIs becomes increasingly
relevant to careers in the software industry. Through this course, we
hope to give you that knowledge by building up from the very basics. In
this chapter, we start by looking at some fundamental concepts around
APIs. We deﬁne what an API is, where it lives, and give a high level
picture of how one is used.
A Frame of Reference
When talking about APIs, a lot of the conversation focuses on abstract
concepts. To anchor ourselves, let’s start with something that is
physical: the server. A server is nothing more than a big computer. It
has all the same parts as the laptop or desktop you use for work, it’s
just faster and more powerful. Typically, servers don’t have a monitor,
keyboard, or mouse, which makes them look unapproachable. The
reality is that IT folks connect to them remotely — think remote
desktop-style — to work on them.
Servers are used for all sorts of things. Some store data; others send
email. The kind people interact with the most are web servers. These
are the servers that give you a web page when you visit a website. To
better understand how that works, here’s a simple analogy:
In the same way that a program like Solitaire waits for
you to click on a card to do something, a web server
runs a program that waits for a person to ask it for a
There’s really nothing magical or spectacular about it. A software
developer writes a program, copies it to a server, and the server runs
the program continuously.
What An API Is and Why It’s Valuable
Websites are designed to cater to people’s strengths. Humans have an
incredible ability to take visual information, combine it with our
experiences to derive meaning, and then act on that meaning. It’s why
you can look at a form on a website and know that the little box with
the phrase “First Name” above it means you are supposed to type in
the word you use to informally identify yourself.
Yet, what happens when you face a very time-intensive task, like
copying the contact info for a thousand customers from one site to
another? You would love to delegate this work to a computer so it can
be done quickly and accurately. Unfortunately, the characteristics that
make websites optimal for humans make them diﬃcult for computers
The solution is an API. An API is the tool that makes a website’s data
digestible for a computer. Through it, a computer can view and edit
data, just like a person can by loading pages and submitting forms.
Figure 1. Communicating with a server.
Making data easier to work with is good because it means people can
write software to automate tedious and labor-intensive tasks. What
might take a human hours to accomplish can take a computer seconds
through an API.
How An API Is Used
When two systems (websites, desktops, smartphones) link up through
an API, we say they are “integrated.” In an integration, you have two
sides, each with a special name. One side we have already talked about:
the server. This is the side that actually provides the API. It helps to
remember that the API is simply another program running on the
server 3. It may be part of the same program that handles web traﬃc,
or it can be a completely separate one. In either case, it is sitting,
waiting for others to ask it for data.
The other side is the “client.” This is a separate program that knows
what data is available through the API and can manipulate it, typically at
the request of a user. A great example is a smartphone app that syncs
with a website. When you push the refresh button your app, it talks to a
server via an API and fetches the newest info.
The same principle applies to websites that are integrated. When one
site pulls in data from the other, the site providing the data is acting as
the server, and the site fetching the data is the client.
Chapter 1 Recap
This chapter focused on providing some foundational terminology and
a mental model of what an API is and how it is used.
The key terms we learned were:
Server: A powerful computer that runs an API
API: The “hidden” portion of a website that is meant for computer
Client: A program that exchanges data with a server through an
Normally, each chapter has a mini homework assignment where you
apply what you learned. Today, however, you get a pass. Go enjoy your
favorite TV show!
In the next chapter, we dive into the mechanics of how a client talks
with an API.
1. David Berlind, ProgrammableWeb’s Directory Hits 10,000 APIs. And
Counting. ProgrammableWeb. September 23, 2013.
2. Adam DuVander, API Growth Doubles in 2010, Social and Mobile are
Trends. ProgrammableWeb. January 3, 2011.
3. Technically, an API is just a set of rules (interface) that the two sides agree
to follow. The company publishing the API then implements their side by
writing a program and putting it on a server. In practice, lumping the
interface in with the implementation is an easier way to think about it.
In Chapter 1, we got our bearings by forming a picture of the two sides
involved in an API, the server and the client. With a solid grasp on the
who, we are ready to look deeper into how these two communicate. For
context, we ﬁrst look at the human model of communication and
compare it to computers. After that, we move on to the speciﬁcs of a
common protocol used in APIs.
Knowing the Rules
People create social etiquette to guide their interactions. One example
is how we talk to each other on the phone. Imagine yourself chatting
with a friend. While they are speaking, you know to be silent. You know
to allow them brief pauses. If they ask a question and then remain
quiet, you know they are expecting a response and it is now your turn
Computers have a similar etiquette, though it goes by the term
“protocol.” A computer protocol is an accepted set of rules that govern
how two computers can speak to each other. Compared to our
standards, however, a computer protocol is extremely rigid. Think for a
moment of the two sentences “My favorite color is blue” and “Blue is
my favorite color.” People are able to break down each sentence and
see that they mean the same thing, despite the words being in diﬀerent
orders. Unfortunately, computers are not that smart.
For two computers to communicate eﬀectively, the server has to know
exactly how the client will arrange its messages. You can think of it like a
person asking for a mailing address. When you ask for the location of a
place, you assume the ﬁrst thing you are told is the street address,
followed by the city, the state, and lastly, the zip code. You also have
certain expectations about each piece of the address, like the fact that
the zip code should only consist of numbers. A similar level of
speciﬁcity is required for a computer protocol to work.
The Protocol of the Web
There is a protocol for just about everything; each one tailored to do
diﬀerent jobs. You may have already heard of some: Bluetooth for
connecting devices, and POP or IMAP for fetching emails.
On the web, the main protocol is the Hyper-Text Transfer Protocol,
better known by its acronym, HTTP. When you type an address like
http://example.com into a web browser, the “http” tells the browser to
use the rules of HTTP when talking with the server.
With the ubiquity of HTTP on the web, many companies choose to
adopt it as the protocol underlying their APIs. One beneﬁt of using a
familiar protocol is that it lowers the learning curve for developers,
which encourages usage of the API. Another beneﬁt is that HTTP has
several features useful in building a good API, as we’ll see later. Right
now, let’s brave the water and take a look at how HTTP works!
Communication in HTTP centers around a concept called the RequestResponse Cycle. The client sends the server a request to do something.
The server, in turn, sends the client a response saying whether or not
the server could do what the client asked.
Figure 1. The Request-Response Cycle.
To make a valid request, the client needs to include four things:
URL (Uniform Resource Locator) 1
List of Headers
That may sound like a lot of details just to pass along a message, but
remember, computers have to be very speciﬁc to communicate with
URLs are familiar to us through our daily use of the web, but have you
ever taken a moment to consider their structure? In HTTP, a URL is a
unique address for a thing (a noun). Which things get addresses is
entirely up to the business running the server. They can make URLs for
web pages, images, or even videos of cute animals.
APIs extend this idea a bit further to include nouns like customers,
products, and tweets. In doing so, URLs become an easy way for the
client to tell the server which thing it wants to interact with. Of course,
APIs also do not call them “things”, but give them the technical name
The request method tells the server what kind of action the client wants
the server to take. In fact, the method is commonly referred to as the
The four methods most commonly seen in APIs are:
GET – Asks the server to retrieve a resource
POST – Asks the server to create a new resource
PUT – Asks the server to edit/update an existing resource
DELETE – Asks the server to delete a resource
Here’s an example to help illustrate these methods. Let’s say there is a
pizza parlor with an API you can use to place orders. You place an order
by making a POST request to the restaurant’s server with your order
details, asking them to create your pizza. As soon as you send the
request, however, you realize you picked the wrong style crust, so you
make a PUT request to change it.
While waiting on your order, you make a bunch of GET requests to
check the status. After an hour of waiting, you decide you’ve had
enough and make a DELETE request to cancel your order.
Headers provide meta-information about a request. They are a simple
list of items like the time the client sent the request and the size of the
Have you ever visited a website on your smartphone that was specially
formatted for mobile devices? That is made possible by an HTTP header
called “User-Agent.” The client uses this header to tell the server what
type of device you are using, and websites smart enough to detect it
can send you the best format for your device.
There are quite a few HTTP headers that clients and servers deal with,
so we will wait to talk about other ones until they are relevant in later
The request body contains the data the client wants to send the server.
Continuing our pizza ordering example above, the body is where the
order details go.
A unique trait about the body is that the client has complete control
over this part of the request. Unlike the method, URL, or headers,
where the HTTP protocol requires a rigid structure, the body allows the
client to send anything it needs.
These four pieces — URL, method, headers, and body — make up a
complete HTTP request.
Figure 2. The structure of an HTTP request.
After the server receives a request from the client, it attempts to fulﬁll
the request and send the client back a response. HTTP responses have
a very similar structure to requests. The main diﬀerence is that instead
of a method and a URL, the response includes a status code. Beyond
that, the response headers and body follow the same format as
Status codes are three-digit numbers that each have a unique meaning.
When used correctly in an API, this little number can communicate a lot
of info to the client. For example, you may have seen this page during
your internet wanderings:
Figure 3. A default 404 web page.
The status code behind this response is 404, which means “Not Found.”
Whenever the client makes a request for a resource that does not exist,
the server responds with a 404 status code to let the client know: “that
resource doesn’t exist, so please don’t ask for it again!”
There is a slew of other statuses in the HTTP protocol, including 200
(“success! that request was good”) to 503 (“our website/API is currently
down.”) We’ll learn a few more of them as they come up in later
After a response is delivered to the client, the Request-Response Cycle
is completed and that round of communication over. It is now up to the
client to initiate any further interactions. The server will not send the
client any more data until it receives a new request.
Figure 4. The structure of an HTTP response.
How APIs Build on HTTP
By now, you can see that HTTP supports a wide range of permutations
to help the client and server talk. So, how does this help us with APIs?
The ﬂexibility of HTTP means that APIs built on it can provide clients
with a lot of business potential. We saw that potential in the pizza
ordering example above. A simple tweak to the request method was
the diﬀerence between telling the server to create a new order or
cancel an existing one. It was easy to turn the desired business
outcome into an instruction the server could understand. Very
This versatility in the HTTP protocol extends to other parts of a request,
too. Some APIs require a particular header, while others require speciﬁc
information inside the request body. Being able to use APIs hinges on
knowing how to make the correct HTTP request to get the result you
Chapter 2 Recap
The goal of this chapter was to give you a basic understanding of HTTP.
The key concept was the Request-Response Cycle, which we broke
down into the following parts:
Request – consists of a URL (http://…), a method (GET, POST, PUT,
DELETE), a list of headers (User-Agent…), and a body (data).
Response – consists of a status code (200, 404…), a list of headers,
and a body.
Throughout the rest of the course, we will revisit these fundamentals as
we discover how APIs rely on them to deliver power and ﬂexibility.
Use the form on the site for Chapter 2 to make the following list of
requests and see what responses you are given.
Send a GET request without any body data.
Send a POST request and type your favorite kind of pizza in the
Send a PUT request and type a new ingredient to add to your
pizza in the body ﬁeld.
Send a DELETE request without any body data.
In the next chapter, we explore what kind of data APIs pass between
the client and the server.
1. The HTTP speciﬁcation actually requires a request to have a URI
(Universal Resource Identiﬁer), of which URLs are a subset, along with URNs
(Uniform Resource Names). We chose URL because it is the acronym readers
already know. The subtle diﬀerences between these three are beyond the
scope of the course.
So far, we’ve learned that HTTP (Hyper-Text Transfer Protocol) is the
underpinning of APIs on the web and that to use them, we need to
know how HTTP works. In this chapter, we explore the data APIs
provide, how it’s formatted, and how HTTP makes it possible.
When sharing data with people, the possibilities for how to display the
information is limited only by human imagination. Recall the pizza
parlor from last chapter — how might they format their menu? It could
be a text-only, bulleted list; it could be a series of photos with captions;
or it could even be only photos, which foreign patrons could point at to
place their order.
A well-designed format is dictated by what makes the
information the easiest for the intended audience to
The same principle applies when sharing data between computers. One
computer has to put the data in a format that the other will
understand. Generally, this means some kind of text format. The most
Notation) and XML (Extensible Markup Language).
Many new APIs have adopted JSON as a format because it’s built on the
web and usable on both the front- and back-end of a web app or
service. JSON is a very simple format that has two pieces: keys and
values. Keys represent an attribute about the object being described. A
pizza order can be an object. It has attributes (keys), such as crust type,
toppings, and order status. These attributes have corresponding values
(thick crust, pepperoni, and out-for-delivery).
Let’s see how this pizza order could look in JSON:
“toppings”: [“cheese”, “pepperoni”, “garlic”],
In the JSON example above, the keys are the words on the left:
toppings, crust, and status. They tell us what attributes the pizza order
contains. The values are the parts to the right. These are the actual
details of the order.
Figure 1. JSON key and value.
If you read a line from left to right, you get a fairly natural English
sentence. Taking the ﬁrst line as an example, we could read it as, “the
crust for this pizza is original style.” The second line can also be read —
in JSON, a value that starts and ends with square brackets () is a list of
values. So, we read the second line of the order as, “the toppings for
this order are: cheese, pepperoni, and garlic.”
Sometimes, you want to use an object as the value for a key. Let’s
extend our pizza order with customer details so you can see what this
might look like:
“toppings”: [“cheese”, “pepperoni”, “garlic”],
In this updated version, we see that a new key, “customer”, is added.
The value for this key is another set of keys and values that provide
details about the customer that placed the order. Cool trick, huh?! This
is called an Associative Array. Don’t let the technical term intimidate you
though – an associative array is just a nested object.
XML has been around since 1996 1. With age, it has become a very
mature and powerful data format. Like JSON, XML provides a few
simple building blocks that API makers use to structure their data. The
main block is called a node.
Let’s see what our pizza order might look like in XML:
XML always starts with a root node, which in our pizza example is
“order.” Inside the order are more “child” nodes. The name of each
node tells us the attribute of the order (like the key in JSON) and the
data inside is the actual detail (like the value in JSON).
Figure 2. XML node and value.
You can also infer English sentences by reading XML. Looking at the line
with “crust”, we could read, “the crust for the pizza is original style.”
Notice how in XML, every item in the list of toppings is wrapped by a
node. You can see how the XML format requires a lot more text to
communicate than JSON does.
How Data Formats Are Used In HTTP
Now that we’ve explored some available data formats, we need to know
how to use them in HTTP. To do so, we will say hello again to one of the
fundamentals of HTTP: headers. In Chapter 2, we learned that headers
are a list of information about a request or response. There is a header
for saying what format the data is in: Content-Type.
When the client sends the Content-Type header in a request, it is telling
the server that the data in the body of the request is formatted a
particular way. If the client wants to send the server JSON data, it will
set the Content-Type to “application/json.” Upon receiving the request
and seeing that Content-Type, the server will ﬁrst check if it
understands that format, and, if so, it will know how to read the data.
Likewise, when the server sends the client a response, it will also set the
Content-Type to tell the client how to read the body of the response.
Sometimes, the client can only speak one data format. If the server
sends back anything other than that format, the client will fail and
throw an error. Fortunately, a second HTTP header comes to the
rescue. The client can set the Accept header to tell the server what data
formats it is able to accept. If the client can only speak JSON, it can set
the Accept header to “application/json.” The server will then send back
its response in JSON. If the server doesn’t support the format the client
requests, it can send back an error to the client to let it know the
request is not going to work.
With these two headers, Content-Type and Accept, the client and server
can work with the data formats they understand and need to work
Figure 3. Data format headers.
Chapter 3 Recap
In this chapter, we learned that for two computers to communicate,
they need to be able to understand the data format passed to them.
We were introduced to 2 common data formats used by APIs, JSON and
XML. We also learned that the Content-Type HTTP header is a useful
way to specify what data format is being sent in a request and the
Accept header speciﬁes the requested format for a response.
The key terms we learned were:
Object: a thing or noun (person, pizza order…)
Key: an attribute about an object (color, toppings…)
Value: the value of an attribute (blue, pepperoni…)
Associative array: a nested object
XML: Extensible Markup Language
Use the form in our site for Chapter 3 to make the following list of
requests and see what responses you are given.
Send a request with: Content-Type header = “application/json”,
Accept header = “application/json”, and data format = “XML”.
Send a request with: Content-Type header = “application/json”,
Accept header = “application/json”, and data format = “JSON”.
Ok, now just try changing things around and seeing what
In the next chapter, we ﬁnd out how two computers can establish trust
using Authentication in order to pass along sensitive data, like customer
details or private content.
Things are starting to pick up in our understanding of APIs. We know
who the client and server are, we know they use HTTP to talk to each
other, and we know they speak in speciﬁc data formats to understand
each other. Knowing how to talk, though, leaves an important question:
how does the server know the client is who it claims to be? In this
chapter, we explore two ways that the client can prove its identity to the
Identities in a Virtual World
You’ve probably registered for an account on a website before. The
process involves the site asking you for some personal information,
most notably a username and a password. These two pieces of
information become your identifying marks. We call these your
credentials. When you visit the website again, you can login by
providing these credentials.
Logging-in with a username and password is one example of a technical
process known as authentication. When you authenticate with a
server, you prove your identity to the server by telling it information
that only you know (at least we hope only you know it). Once the server
knows who you are, it can trust you and divulge the private data in your
There are several techniques APIs use to authenticate a client. These
are called authentication schemes. Let’s take a look at two of these
The logging-in example above is the most basic form of authentication.
In fact, the oﬃcial name for it is Basic Authentication (“Basic Auth” to
its friends). Though the name has not garnered any creativity awards,
the scheme is a perfectly acceptable way for the server to authenticate
the client in an API.
Basic Auth only requires a username and password. The client takes
these two credentials, smooshes them together to form a single value 1,
and passes that along in the request in an HTTP header called
Figure 1. The Authorization HTTP header.
When the server receives the request, it looks at the Authorization
header and compares it to the credentials it has stored. If the username
and password match one of the users in the server’s list, the server
fulﬁlls the client’s request as that user. If there is no match, the server
returns a special status code (401) to let the client know that
authentication failed and the request is denied.
Though Basic Auth is a valid authentication scheme, the fact that it uses
same username and password to access the API and manage the
account is not ideal. That is like a hotel handing a guest the keys to the
whole building rather than to a room.
Similarly with APIs, there may be times when the client should have
diﬀerent permissions than the account owner. Take for example a
business owner who hires a contractor to write a program that uses an
API on their behalf. Trusting the contractor with the account credentials
puts the owner at risk because an unscrupulous contractor could
change the password, locking the business owner out of their own
account. Clearly, it would be nice to have an alternative.
API Key Authentication
API Key authentication is a technique that overcomes the weakness of
using shared credentials by requiring the API to be accessed with a
unique key. In this scheme, the key is usually a long series of letters and
numbers that is distinct from the account owner’s login password. The
owner gives the key to the client, very much like a hotel gives a guest a
key to a single room.
When the client authenticates with the API key, the server knows to
allow the client access to data, but now has the option to limit
administrative functions, like changing passwords or deleting accounts.
Sometimes, keys are used simply so the user does not have to give out
their password. The ﬂexibility is there with API Key authentication to
limit control as well as protect user passwords.
So, where does the API key go? There is a header for it, too, right?
Unfortunately, there is not. Unlike Basic Auth, which is an established
standard with strict rules, API keys were conceived at multiple
companies in the early days of the web. As a result, API key
authentication is a bit like the wild west; everybody has their own way
of doing it.
Over time, however, a few common approaches have emerged. One is
to have the client put the key in the Authorization header, in lieu of a
username and password. Another is to add the key onto the URL
(http://example.com?api_key=my_secret_key). Less common is to bury
the key somewhere in the request body next to the data. Wherever the
key goes, the eﬀect is the same – it lets the server authenticate the
Chapter 4 Recap
In this chapter, we learned how the client can prove its identity to the
server, a process known as authentication. We looked at two
techniques, or schemes, APIs use to authenticate.
The key terms we learned were:
Authentication: process of the client proving its identity to the
Credentials: secret pieces of info used to prove the client’s
identity (username, password…)
Basic Auth: scheme that uses an encoded username and
password for credentials
API Key Auth: scheme that uses a unique key for credentials
Authorization Header: the HTTP header used to hold credentials
Use the form in our site for Chapter 4 below to explore locations using
the Google Maps API.
In the next chapter, we continue the discussion of authentication by
looking at a third technique; one that is quickly becoming the standard
of the web.
1. The actual process involves combining the username with a colon,
followed by the password, and then running the whole string through the
base64 encoding algorithm. Thus “user” and “password” becomes
“user:password” and, after encoding, you have “dXNlcjpwYXNzd29yZAo=”.
In Chapter 4, we mentioned most websites use a username and
password for authentication credentials. We also discussed how reusing
these credentials for API access is insecure, so APIs often require a
diﬀerent set of credentials from the ones used to login to the website. A
common example is API keys. In this chapter, we look at another
solution, Open Authorization (OAuth), which is becoming the most
widely used authentication scheme on the web.
Making Life Easy for People
Have you ever had to complete a registration form like the one below?
Figure 1. A product key as seen on Microsoft’s Windows 8 registration form.
Typing a long key into a form ﬁeld like the one above makes for a poor
user-experience. First, you have to ﬁnd the required the key. Sure, it
was right in your inbox when you bought the software, but a year later,
you’re scrambling to ﬁnd it (What email was it sent from? Which email
did I use to register?!) Once located, you have to enter the darned thing
perfectly – making a typo or missing a single character will result in
failure, or might even get you locked out of your unregistered software!
Forcing users to work with API keys is a similarly poor experience. Typos
are a common problem and it requires the user to do part of the setup
between the client and server manually. The user must obtain the key
from the server, then give it to the client. For tools meant to automate
work, surely there’s a better solution.
Enter OAuth. Automating the key exchange is one of the main problems
OAuth solves. It provides a standard way for the client to get a key from
the server by walking the user through a simple set of steps. From the
user’s perspective, all OAuth requires is entering credentials. Behind
the scenes, the client and server are chattering back and forth to get
the client a valid key.
There are currently two versions of OAuth, aptly named OAuth 1 and
OAuth 2. Understanding the steps in each is necessary to be able to
interact with APIs that use them for authentication. Since they share a
common workﬂow, we will walk through the steps of OAuth 2, then
point out the ways in which OAuth 1 diﬀers.
To get started, we ﬁrst need to know the cast of characters involved in
an OAuth exchange:
The User – A person who wants to connect two websites they use
The Client – The website that will be granted access to the user’s
The Server – The website that has the user’s data
Next, we need to give a quick disclaimer. One goal of OAuth 2 is to allow
businesses to adapt the authentication process to their needs. Due to
this extendable nature, APIs can have slightly diﬀerent steps. The
workﬂow shown below is a common one found among web-based
apps. Mobile and desktop applications might use slight variations on
With that, here are the steps of OAuth 2.
Step 1 – User Tells Client to Connect to Server
The user kicks oﬀ the process by letting the client know they want it to
connect to the server. Usually, this is by clicking a button.
Step 2 – Client Directs User to Server
The client sends the user over to the server’s website, along with a URL
that the server will send the user back to once the user authenticates,
called the callback URL.
Step 3 – User Logs-in to Server and Grants Client
With their normal username and password, the user authenticates with
the server. The server is now certain that one of its own users is
requesting that the client be given access to the user’s account and
Step 4 – Server Sends User Back to Client, Along with
The server sends the user back to the client (to the Callback URL from
Step 2). Hidden in the response is a unique authorization code for the
Step 5 – Client Exchanges Code + Secret Key for Access
The client takes the authorization code it receives and makes another
request to the server. This request includes the client’s secret key.
When the server sees a valid authorization code and a trusted client
secret key, it is certain that the client is who it claims to be and that it is
acting on behalf of a real user. The server responds back with an access
Step 6 – Client Fetches Data from Server
At this point, the client is free to access the server on the user’s behalf.
The access token from Step 6 is essentially another password into the
user’s account on the server. The client includes the access token with
every request so it can authenticate directly with the server.
Client Refreshes Token (Optional)
A feature introduced in OAuth 2 is the option to have access tokens
expire. This is helpful in protecting users’ accounts by strengthening
security – the faster a token expires, the less time a stolen token might
be used maliciously, similar to how a credit card number expires after a
certain time. The lifespan of a token is set by the server. APIs in the wild
use anything from hours to months. Once the lifespan is reached, the
client must ask the server for a new token.
How OAuth 1 Is Diﬀerent
There are several key diﬀerences between the versions of OAuth. One
we already mentioned; access tokens do not expire.
Another distinction is that OAuth 1 includes an extra step. Between
Steps 1 and 2 above, OAuth 1 requires the client to ask the server for a
request token. This token acts like the authorization code in Oauth 2
and is what gets exchanged for the access token.
A third diﬀerence is that OAuth 1 requires requests to be digitally
signed. We’ll skip the details of how signing works (you can ﬁnd code
libraries to do this for you), but it is worth knowing why it is in one
version and not the other. Request signing is a way to protect data from
being tampered with while it moves between the client and the server.
Signatures allow the server to verify the authenticity of the requests.
Today, however, most API traﬃc happens over a channel that is already
secure (HTTPS). Recognizing this, OAuth 2 eliminates signatures in an
eﬀort to make version two easier to use. The trade-oﬀ is that OAuth 2
relies on other measures to provide security to the data in transit.
An element of OAuth 2 that deserves special attention is the concept
limiting access, known formally as authorization. Back in Step 2, when
the user clicks the button to allow the client access, buried in the ﬁne
print are the exact permissions the client is asking for. Those
permissions, called scope, are another important feature of OAuth 2.
They provide a way for the client to request limited access to the user’s
data, thereby making it easier for the user to trust the client.
What makes scope powerful is that it is client-based restrictions. Unlike
an API Key, where limits placed on the key aﬀect every client equally,
OAuth scope allows one client to have permission X and another
permissions X and Y. That means one website might be able to view
your contacts, while another site can view and edit them.
Chapter 5 Recap
In this chapter, we learned the ﬂow of the OAuth authentication
process. We compared the two versions, pointing out the major
diﬀerence between them.
The key terms we learned were:
OAuth: an authentication scheme that automates the key
exchange between client and server.
Access Token: a secret that the client obtains upon successfully
completing the OAuth process.
Scope: permissions that determine what access the client has to
This exercise walks you through the OAuth 2 ﬂow for Facebook. In this
scenario, you play the role of the user, we are the client, and Facebook
is the server.
You will connect your real account and have us pull a few friends from your
list to illustrate how the process works.
We will not store any of the data we get, nor do we keep the access
token. As soon as you leave this page, we will disconnect with Facebook
and all temporary data will be removed.
Visit our site for Chapter 5 to run the exercise.
In the next chapter, we look at some of the basics concepts in API
design, including how APIs organize their data so the client can easily
access what it wants.
This chapter marks a turning point in our adventure with APIs. We are
ﬁnished covering fundamentals and are now ready to see how the
previous concepts combine to form an API. In this chapter, we discuss
the components of an API by designing one.
National Geographic estimated that in 2011, Americans snapped 80
billion photos 1. With so many photos, you can imagine the diﬀerent
approaches people have to organizing them on their computers. Some
people prefer to dump everything into a single folder. Others
meticulously arrange their pictures into a hierarchy of folders by year,
month, and event.
Companies give similar thought to organization when building their
APIs. As we mentioned in Chapter 1, the purpose of an API is to make it
easy for computers to work with the company’s data. With ease of use
in mind, one company may decide to have a single URL for all the data
and make it searchable (sort of like having one folder for all your
photos). Another may decide to give each piece of data its own URL,
organized in a hierarchy (like having folders and sub-folders for
photos). Each company chooses the best way to structure its API for its
particular situation, guided by existing industry best practices.
Start with an Architectural Style
When discussing APIs, you might hear talk of “soap” and “rest” and
wonder whether the software developers are doing work or planning a
vacation. The truth is that these are the names of the two most
common architectures for web-based APIs. SOAP (formerly an acronym
2) is an XML-based design that has standardized structures for requests
and responses. REST, which stands for Representational State Transfer,
is a more open approach, providing lots of conventions, but leaving
many decisions to the person designing the API.
Throughout this course, you may have noticed we’ve had an inclination
for REST APIs. The preference is largely due to REST’s incredible rate of
adoption 3. This is not to say that SOAP is evil; it has its strong points 4.
However, the focus of our discussion will stay on REST as this will likely
be the kind of API you encounter. In the remaining sections, we walk
through the components that make up a REST API.
Our First Resource
Back in Chapter 2, we talked a little bit about resources. Recall that
resources are the nouns of APIs (customers and pizzas). These are the
things we want the world to be able to interact with through our API.
To get a feel for how a company would design an API, let’s try our hand
at it with our pizza parlor. We’ll start by adding the ability to order a
For the client to be able to talk pizzas with us, we need to do several
Decide what resource(s) need to be available.
Assign URLs to those resources.
Decide what actions the client should be allowed to perform on
Figure out what pieces of data are required for each action and
what format they should be in.
Picking resources can be a diﬃcult ﬁrst task. One way to approach the
problem is to step through what a typical interaction involves. For our
pizza parlor, we probably have a menu. On that menu are pizzas. When
a customer wants us to make one of the pizzas for them, they place an
order. In this context, menu, pizza, customer, and order all sound like
good candidates for resources. Let’s start with order.
The next step is assigning URLs to the resource. There are lots of
possibilities, but luckily REST conventions give some guidance. In a
typical REST API, a resource will have two URL patterns assigned to it.
The ﬁrst is the plural of the resource name, like /orders. The second is
the plural of the resource name plus a unique identiﬁer to specify a
single resource, like /orders/<order_id>, where <order_id> is the
unique identiﬁer for an order. These two URL patterns make up the ﬁrst
endpoints that our API will support. These are called endpoints simply
because they go at the end of the URL, as in http://example.com/
Now that we picked our resource and assigned it URLs, we need to
decide what actions the client can perform. Following REST
conventions, we say that the plural endpoint (/orders) is for listing
existing orders and creating new ones. The plural with a unique
identiﬁer endpoint (/orders/<order_id>), is for retrieving, updating, or
cancelling a speciﬁc order. The client tells the server which action to
perform by passing the appropriate HTTP verb (GET, POST, PUT or
DELETE) in the request.
Altogether, our API now looks like this:
List existing orders
Place a new order
Get details for order #1
Get details for order #2
Update order #1
Cancel order #1
With the actions for our order endpoints ﬂeshed out, the ﬁnal step is to
decide what data needs to be exchanged between the client and the
server. Borrowing from our pizza parlor example in Chapter 3, we can
say that an order needs a crust and toppings. We also need to select a
data format that the client and server can use to pass this information
back and forth. XML and JSON are both good choices, but for readability
sake, we’ll go with JSON.
At this point, you should pat yourself on the back; we have designed a
functional API! Here is what an interaction between the client and
server might look like using this API:
Figure 1. Example interaction between the client and server using our API.
Linking Resources Together
Our pizza parlor API is looking sharp. Orders are coming in like never
before. Business is so good in fact, we decide we want to start tracking
orders by customer to gauge loyalty. An easy way to do this is to add a
new customer resource.
Just like with orders, our customer resource needs some endpoints.
Following convention, /customers and /customers/<id> ﬁt nicely. We’ll
skip the details, but let’s say we decide which actions make sense for
each endpoint and what data represents a customer. Assuming we do
all of that, we come to an interesting question: how do we associate
orders with customers?
REST practitioners are split on how to solve the problem of associating
resources. Some say that the hierarchy should continue to grow, giving
endpoints like /customers/5/orders for all of customer #5’s orders
and /customers/5/orders/3 for customer #5’s third order. Others argue
to keep things ﬂat by including associated details in the data for a
resource. Under this paradigm, creating an order requires a
customer_id ﬁeld to be sent with the order details. Both solutions are
used by REST APIs in the wild, so it is worth knowing about each.
Figure 2. Two ways to handle associated data in API design.
As data in a system grows, endpoints that list all records become
impractical. Imagine if our pizza parlor had three million completed
orders and you wanted to ﬁnd out how many had pepperoni as a
topping. Sending a GET request to /orders and receiving all three
million orders would not be very helpful. Thankfully, REST has a nifty
way for searching through data.
URLs have another component that we have not mentioned yet, the
query string. Query means search and string means text. The query
string is a bit of text that goes onto the end of a URL to pass things
along to the API. For example, everything after the question mark is the
query string in http://example.com/orders?key=value.
REST APIs use the query string to deﬁne details of a search. These
details are called query parameters. The API dictates what parameters
it will accept, and the exact names of those parameters need to be used
for them to eﬀect the search. Our pizza parlor API could allow the client
to search for orders by topping by using this URL: http://example.com/
orders?topping=pepperoni. The client can include multiple query
parameters by listing one after another, separating them by an
ampersand (“&”). For example: http://example.com/orders?
Another use of the query string is to limit the amount of data returned
in each request. Often, APIs will split results into sets (say of 100 or 500
records) and return one set at a time. This process of splitting up the
data is known as pagination (an analogy to breaking up words into
pages for books). To allow the client to page through all the data, the
API will support query parameters that allow the client to specify which
page of data it wants. In our pizza parlor API, we can support paging by
allowing the client to specify two parameters: page and size. If the client
makes a request like GET /orders?page=2&size=200, we know they want
the second page of results, with 200 results per page, so orders
Chapter 6 Recap
In this chapter, we learned how to design a REST API. We showed the
basic functions an API supports and how to organize the data so that it
can be easily consumed by a computer.
The key terms we learned were:
SOAP: API architecture known for standardized message formats
REST: API architecture that centers around manipulating
Resource: API term for a business noun like customer or order
Endpoint: A URL that makes up part of an API. In REST, each
resource gets its own endpoints
Query String: A portion of the URL that is used to pass data to the
Query Parameters: A key-value pair found in the query string
Pagination: Process of splitting up results into manageable
Your homework for this chapter is an exploration of API design. We’ll
look at a few examples using 3 notable APIs to see what’s available and
how things are structured.
Example 1: The Instagram API
Answer the following questions about Instagram’s API design.
To ﬁnd the answer to the following 3 questions, open the Instagram API
1. What resources does Instagram make available (hint: look at the
2. What is unique identiﬁer for users?
3. For the endpoint users/self/media/liked, what is the name of
the parameter that limits the number of media results returned?
Example 2: The Facebook API
Answer the following questions about Facebooks’s API design.
1. What 3 terms does Facebook use to describe what the Graph API
is composed of?
Open “Quickstart” to ﬁnd the answer.
2. What does ‘me’ in the /me endpoint translate to as a convenience?
Open “Using the Graph API” to ﬁnd the answer.
Example 3: The Twitter API
Answer the following questions about Twitter’s API design.
1. What 4 resources, referred to as “objects”, does Twitter make
Open the docs index to ﬁnd the answer.
2. What parameter is required to create a new favorite?
Open the “POST favorites/create” to ﬁnd the answer.
In the next chapter, we explore ways to make the client react to
changes on the server in real-time.
1. Unknown, Image Obsessed. National Geographic. April, 2012.
2. SOAP stood for Simple Object Access Protocol. It was originally used for a
very speciﬁc type of API access. As developers found ways to apply it to more
situations, the name no longer ﬁt, so in SOAP version 1.2 the acronym was
3. Abel Avram, Is REST Successful in the Enterprise?. InfoQ. June 1, 2011.
4. SOAP provides a very structured architecture. The structure provides
system reliability, standard extensions for adding functionality to the
protocol, and makes it possible for tools to generate code, saving on
In Chapter 6, we learned about designing APIs by building our own and
exploring a few real-world examples. At this point, we have a lot of
hard-earned knowledge and it’s time for it to start paying oﬀ. We are
ready to see how we can put APIs to work for us. In this chapter, we
learn four ways to achieve real-time communication through APIs.
To set the stage for our discussion, let’s remind ourselves why APIs are
useful. Back in Chapter 1 we said that APIs make it easy to share data
between two systems (websites, desktops, smartphones).
Straightforward sharing allows us to link systems together to form an
integration. People like integrations because they make life easier. With
an integration, you can do something in one system and the other will
For our purposes, we will split integrations into two broad categories1.
The ﬁrst we call “client-driven,” where a person interacts with the client
and wants the server’s data to update. The other we call “serverdriven”, where a person does something on the server and needs the
client to be aware of the change.
The reason for dividing integrations in this manner comes down to one
simple fact: the client is the only one who can initiate communication.
Remember, the client makes requests and the server just responds. A
consequence of this limitation is that changes are easy to send from the
client to the server, but hard to do in the reverse direction.
To demonstrate why client-driven integrations are easy, let’s turn to our
trusty pizza parlor and its API for ordering pizzas. Say we release a
smartphone app that uses the API. In this scenario, the pizza parlor API
is the server and the smartphone app is the client. A customer uses the
app to choose a pizza and then hits a button to place the order. As soon
as the button is pressed, the app knows it needs to make a request to
the pizza parlor API.
Figure 1. Example of a client-driven interaction.
More generally speaking, when a person interacts with the client, the
client knows exactly when data changes, so it can call the API
immediately to let the server know. There’s no delay (hence it’s realtime) and the process is eﬃcient because only one request is made for
each action taken by a person.
Once the pizza order is placed, the customer might want to know when
the pizza is ready. How do we use the API to provide them with
updates? Well, that is a bit harder. The customer has nothing to do with
making the pizza. They are waiting on the pizza parlor to prepare the
pizza and update the order status. In other words, the data is changing
on the server and the client needs to know about it. Yet, if server can’t
make requests, we appear to be stuck!
Solving this type of problem is where we utilize the second category of
integrations. There are a number of solutions software developers use
to get around the client-only requests limitation. Let’s take a look at
When the client is the only one who can make requests, the simplest
solution to keep it up-to-date with the server is for the client to simply
ask the server for updates. This can be accomplished by repeatedly
requesting the same resource, a technique known as polling.
With our pizza parlor, polling for the status of an order might look like
Figure 2. Example of polling for the status of an order in our Pizza Parlor.
In this approach, the more frequently the client makes requests (polls),
the closer the client gets to real-time communication. If the client polls
every hour, at worst, there could be a one-hour delay between a
change happening on the server and the client becoming aware of it.
Poll every minute instead and the client and server eﬀectively stay in
Of course, there is one big ﬂaw with this solution. It is terribly
ineﬃcient. Most of the requests the client makes are wasted because
nothing has changed. Worse, to get updates sooner, the polling interval
has to be shortened, causing the client to make more requests and
become even more ineﬃcient. This solution does not scale well.
If requests were free, then nobody would care about eﬃciency and
everyone could just use polling. Unfortunately, handling requests
comes at a cost. For an API to handle more requests, it needs to utilize
more servers, which costs more money. Scale this cumbersome
situation up to Google- or Facebook-sized proportions, and you’re
paying a lot for ineﬃciency. Hence, lots of eﬀort has been put into
optimizing the way the client can receive updates from the server.
One optimization, which builds oﬀ of polling, is called long polling. Long
polling uses the same idea of the client repeatedly asking the server for
updates, but with a twist: the server does not respond immediately.
Instead, the server waits until something changes, then responds with
Let’s revisit the polling example from above, but this time with a server
that uses the long polling trick.
Figure 3. Long polling example.
This technique is pretty clever. It obeys the rule of the client making the
initial request while leveraging the fact that there is no rule against the
server being slow to respond. As long as both the client and the server
agree that the server will hold on to the client’s request, and the client
is able to keep its connection to the server open, it will work.
As resourceful as long polling is, it too has some drawbacks. We’ll skip
the technical details, but there are concerns like how many requests
can the server hold on to at a time, or how to recover if the client or
server loses its connection. For now, we’ll say that for some scenarios,
neither form of polling is suﬃcient.
With polling ruled out, some innovative software developers thought, “if
all our trouble is because the client is the only one making requests,
why not remove that rule?” So they did. The result was webhooks, a
technique where the client both makes requests and listens for them,
allowing the server to easily push updates to it.
If this sounds like cheating because now we have the server making
requests to the client, don’t worry, you’ve not been told a lie. What
makes webhooks work is that the client becomes a server too! From a
technical perspective, it’s sometimes very easy to extend the client’s
functionality to also listen for requests, enabling two-way
Let’s look at the basics of webhooks. In its simplest form, webhooks
requires the client to provide a Callback URL where it can receive
events, and the server to have a place for a person to enter that
Callback URL. Then, whenever something changes on the server, the
server can send a request to the client’s Callback URL to let the client
know about the update.
For our pizza parlor, the ﬂow might look a little something like the
Figure 4. Using Webhooks to Receive Updates (with Zapier as the client).
This solution is excellent. Changes happening on the server are sent
instantly to the client, so you have true real-time communication. Also,
webhooks are eﬃcient since there is only one request per update.
Building on the idea of webhooks, there have been a variety of
solutions that aim to make the setup process dynamic and not require
a person to manually enter a Callback URL on the server. You might
hear names like HTTP Subscriptions Speciﬁcation, Restful Webhooks,
REST Hooks, and PubSubHubbub. What all of these solutions try to do
is deﬁne a subscription process, where the client can tell the server
what events it is interested in and what Callback URL to send updates
Each solution has a slightly diﬀerent take on the problem, but the
general ﬂow looks like the following.
Figure 5. Requests needed for Subscriptions Webhooks.
Subscription-based webhooks hold a lot of promise. They are eﬃcient,
real-time, and easy for people to use. Similar to REST’s explosive
adoption, a tide is rising behind the movement and it is becoming more
common for APIs to support some form of webhooks.
Still, there will likely be a place for polling and long polling for the
foreseeable future. Not all clients can also act as servers. Smartphones
are a great example where technical constraints rule out webhooks as a
possibility. As technology progresses, new ideas will emerge for how to
make real-time communication easier between all kinds of devices.
Chapter 7 Recap
In this chapter, we grouped integrations into two broad categories,
client-driven and server-driven. We saw how APIs can be used to
provide real-time updates between two systems, as well as some of the
The key terms we learned were:
Polling: Repeatedly requesting a resource at a short interval
Long Polling: Polling, but with a delayed response; improves
Webhooks: When the client gives the server a Callback URL, so the
server can post updates in real time
Subscription Webhooks: Informal name for solutions that make
setting up webhooks automatic
Your homework for this chapter is to play with a little app that
introduces the dimension of time to give you a feel for the diﬀerent
Check out our site for Chapter 7 to try the app.
In the ﬁnal chapter of this course, we look at what it takes to turn an API
design into working software.
1. Client-driven and server-driven are our terms, so don’t be surprised if you
use one in front of a developer and get only a blank stare in return. Mention
polling or webhooks if you want instant credibility.
We made it! We now know everything there is to know about APIs…at
an introductory level at least. So, with all this acquired knowledge, how
can we put it to good use? In this chapter, we explore how to turn
knowledge into working software.
From Plan to Product
As we have seen throughout this course, an API interaction involves two
sides. When we are talking at the code-level, though, what we are really
saying is that we need two programs that implement the API. A
program implements an API when it follows the API’s rules. In our pizza
parlor example, a client that can make requests to the /orders endpoint
using the correct headers and data format would be a client that
implements the pizza parlor’s API.
The server program is the responsibility of the company publishing the
API. Back in Chapter 6, we looked at the process behind designing the
API. After planning, the next step is for the company to implement their
side by writing software that follows the design. The last step is to put
the resulting program on a server.
Along with the server software, the company publishes documentation
for the API. The documentation is one or more documents — typically
webpages or PDFs — that explain how to use the API. It includes
information like what authentication scheme to use, what endpoints
are available, and how the data is formatted. It may also include
example responses, code snippets, and an interactive console to play
with the available endpoints. Documentation is important because it
acts as a guide for building clients. It’s where someone interested in
using the API goes to learn how the API works.
With documentation in hand, there are a number of ways you can begin
to use an API as a client. Let’s examine three of those now.
An easy way to start using an API is with an HTTP Client, a generic
program that lets you quickly build HTTP requests to test with. You
specify the URL, headers, and body, and the program sends it to the
server properly formatted. These types of programs come in all sorts of
ﬂavors, including web apps, desktop apps, web browser extensions,
Figure 1. Screenshot of Dev HTTP Client, a Google Chrome Extension.
The nice thing about generic HTTP clients is that you do not have to
know how to program to use one. With the skills you’ve attained
through this course, you now have the ability to read a company’s API
documentation and ﬁgure out the request you need to make to get the
data you want. This small learning curve makes generic clients great for
exploration and quick one-oﬀ tasks.
There are couple of downsides to this approach, however. First, you
usually can’t save your work. After you close the program, the requests
you made are forgotten and you have to rebuild them the next time
you need them. Another disadvantage is that you typically can’t do
much with the data you get back, other than look at it. At best, you have
the ability to save the data into a ﬁle, after which it’s up to you to do
something interesting with it.
To really harness the power of an API, you will eventually need custom
software. This is where programming comes in. Being a discipline unto
itself, we won’t attempt to cover everything about software
development, but we can give you some guidance for what writing an
API client involves.
The ﬁrst requirement is to gain some familiarity with a programming
language. There are a bunch out there, each with its strengths and
weaknesses. For simplicity’s sake, it is probably better that you stick to
instead of a compiled language (C or C++).
If you aren’t sure which language to choose, a great way to narrow
down the selection can be to ﬁnd an API you want to implement and
see if the company provides a client library. A library is code that the
API owner publishes that already implements the client side of their
API. Sometimes the library will be individually available for download or
it will be bundled in an SDK (Software Development Kit). Using a library
saves you time because instead of reading the API documentation and
forming raw HTTP requests, you can simply copy and paste a few lines
of code and already have a working client.
After you settle on a language, you need to decide where the code will
run. If you are automating your own tasks, running the software from
your work computer might be acceptable. More frequently, you will
want to run the code on a computer better suited for acting as a web
server. There are quite a few solutions available, including running your
code on shared hosting environments, cloud services (like Amazon Web
Services), or even on your own physical servers at a data center.
A third important decision is to ﬁgure out what you will do with the
data. Saving results into a ﬁle is easy enough, but if you want to store
the data in a database or send it to another application, things become
more complex. Pulling data out of a database to send to an API can also
At this point we can pause and remind you to not be too intimidated by
all this new information. You should not expect to know everything
about implementing APIs on your ﬁrst attempt. Take solace knowing
that there are people who can help (open source communities,
developers for-hire, and potential project collaborators) and lots of
resources available online to facilitate learning.
Once you master the basics, there are plenty more topics to learn about
in the rich realm of software development. For now, if you succeed at
learning a programming language and getting a library up and running,
you should celebrate. You will be well on your way to making the most
Give Zapier A Try
If coding is beyond your current skill set or time constraints, there is a
nifty tool we know of that empowers you to easily interact with APIs.
OK, you probably saw this coming: it’s Zapier! Our Developer Platform
oﬀers a way for you to implement an API that you then interact with as
an app on Zapier. Through button clicks and ﬁling out forms, you can
implement nearly any API you want.
What makes using the Developer Platform easy is that we have done a
lot of the programming for you. Zapier has code in place to make
requests, all you have to do is ﬁll in the details. Think of using the
platform a bit like using a generic HTTP Client; you tell us a bit about
the endpoints and we’ll do the rest. The additional beneﬁt is that once
you have Zapier talking with an API, you have lots of options for what to
do with the data you get back. Also, if you get stuck, you are welcome to
reach out to the friendly support team, where you have API experts
ready to help you out.
We’ve come a long way in this course. We started with questions like
“What is a client?” and end with ideas for how to build them. Even if you
decide not to try your hand at implementing an API, we hope you feel
comfortable when they come up in conversation. You now have a sense
of what an API is, what it does, and how you can leverage one to beneﬁt
Maybe you run a company and see the value in providing an API to your
customers. Or perhaps you regularly do pesky, time-consuming tasks
that you are excited to have automated. Whatever the case may be, we
hope you ﬁnd the knowledge you’ve gained valuable. Please share this
course with anyone you think could beneﬁt from it and spread the word
that APIs are awesome!
Chapter 8 Recap
In this chapter, we discussed how an API design becomes working
software. We talked about ways that you can begin using APIs.
The key terms we learned were:
Implement: Writing software that obeys the rules of an API
Documentation: Webpages, PDF’s, etc. that explain the rules of an
Library: Code released by an API publisher that implements the
client portion of their API
Think about ways you might be able to use an API in your working life.
To get the juices going, here are a few ideas:
You need some quick stats from a SaaS (Software as a Service)
application you use. Firing up an HTTP Client to make a few
requests could be a fast way to get the information you need.
You have a labor-intensive task that needs to get done and there
isn’t time to have a developer friend lend a hand. Grabbing a
client library and creating a quick program could be a big
You really want to move data between two SaaS apps on a
continual basis, but you don’t have the resources to build a client
for each app from scratch, nor a good place to run that code.
Using the Zapier Developer Platform could be a low-cost way to
get the applications connected.