Application programming interfaces (APIs), in the most basic sense, simply allow one software or service to interact with another. While APIs are a very valuable business tool, they can be complicated to build and there aren’t a lot of them readily available. Kimono is tackling the issue with its browser-based tool that simplifies the creation of an API. The startup, which recently graduated from the Y Combinator accelerator program, launched in January and has big plans to shake up the web.
According to Kimono co-founder Pratap Ranade, the company is setting out to solve two problems.
“There’s two things going on. One is, there’s a ton of data on the web, most of what you need to find is there; but, it’s massively unstructured and it’s really painful for the developer to go get the data. The second thing is there’s a group of people who don’t even know about web scraping, or can’t even write the web scrapers, but they’re actually really smart and know what questions to ask of the data,” said Ranade.
Ranade met co-founder Ryan Rowe when they both dropped out of Ph.D. programs at Columbia University in New York. Ranade went on to work for McKinsey & Company in New York while Rowe went on to work for Frog Design in their Shanghai studio. When they crossed paths, the two would hack projects together for fun. As they were working on an app that could tell you what movies were playing on your next flight, they hit their pain point — writing custom data scrapers was difficult and tedious, and it often took a long time.
So, they set out to build an API tool for themselves that was easy enough for anyone to use, and Kimono was born.
Building an API
Kimono’s browser tool gives people without a coding background the ability to glean specific data points off of the websites of their choosing and turn that data into a mobile app or import it into another program.
“The way it works is there’s a bookmarklet that you just add to your bookmarks bar,” Ranade said. “When you click it, we kind of take over the page and the whole thing becomes selectable. When you make your first click, Kimono recognizes other elements that are structurally similar to what clicked on and suggest them to you.”
As you click, the product builds a data model in the background. It figures out where you are clicking and tries to determine a common pattern in the data you are clicking on. It sees that you are clicking on specific fields (e.g. titles, authors names, etc.), then it sees groups of data that share the same properties as what you clicked on. When you click done, it creates an API.
The API is basically a big data feed of all the data points you chose on that specific website. The API is set on a schedule to periodically revisit that site and scrape any new data. After it retrieves the data and structures it, the API types it into whatever application you have decided to build. So, if your app is contingent on house prices in a certain zip code, the API will update that for you as it fluctuates so you don’t have to manually go in and update it yourself. Keep in mind this is unstructured data, so certain changes in the website structure or design could potentially affect the API.
“Kimono not only replaces cumbersome processes but opens up entirely new ones that would have been very difficult to do otherwise. At Silk.co, we are building many many databases/visualizations of high-interest web topics. So, using something like Kimono speeds up our ability to produce the databases (probably by as much as 10x) and helps us scale our content production operations,” said Alex Salkever, head of product marketing at Silk.co.
This timelapse map of when and where the Sochi 2014 olympians were born was created by Javier de la Torre.
Creating something cool
The company launched in January, so they are currently running in beta mode and still finding their legs. Regardless, they still have some interesting use cases. In the early days many of the projects were built around Bitcoin. The Kimono team even took to building a Bitcoin Correlator that allowed users to click on any piece of data on the internet and see how Bitcoin prices correlate to those individual pieces of data (e.g. how Bitcoin prices correlate to gold prices). They also built an API for the Sochi Olympics that allowed users to enter queries about personal data of the athletes and the history of the teams.
Luciano Caratori, an energy analyst in Argentina, used Kimono to create a data feed that tracks hourly energy demand and generation costs for the national power grid in Argentina, which he feeds to his Twitter account Potsadi.
“Potsadi would not exist without this service. Since I have no web scraping skills, and this was a personal project born to address a specific need, the existence of this bot was contingent on finding a friendly, powerful, and cost-efficient tool such as Kimono,” said Caratori.
According to Ranade, Kimono wants to empower users to experiment, and that is part of how they plan to monetize the service. Like many founders, they are conflicted by the intersection of the desire to empower their colleagues to create cool stuff and the desire to turn this into a real business. At this point the tool is free, and they hope to keep it free to use for developers who are looking to experiment, but they might end up charging bigger businesses, or companies that are using APIs to create sellable products.
“Our mission is to API-ify the whole internet, make that whole internet machine-readable,” Ranade said.
Moving forward, Kimono wants to focus on getting the tool right and building it out to be more robust.