Note: most of this blog post is centered around the internal debate regarding whether we should make our API public or not. We almost forgot that we announce major feature updates and releases here on the blog. Perils of having a busy life and growing faster than you can handle. Well, we now have the Zomato API finally released for developers to integrate Zomato content in their applications. We had been receiving quite a lot of requests from the developer community for almost an year now, and trust me when I say that building and rolling this out was one helluva difficult job. For two primary reasons:

The API had to be robust

Most of our website code is not robust; we barely survive on an everyday basis.. there are bug reports and crashes and fixes everyday, we have learnt to live with them and after so many years, things are finally starting to look better. We could not do this to our API; since an API very quickly becomes a standard and when people other than your own team start using your code, you have very little flexibility to change stuff realtime. We started developing the API around an year ago – tested it out with our own mobile apps – our Android app which was out in January 2011 was running on the same API. Some partners who trusted our team (Makemytrip, Mapmyindia – thankyou!) helped us take the API live with their own applications and had us test it for scalability. Net result – our API is now very fast and can handle 10,000 requests a second without blinking a processor core! All our apps are built using the same API (of course, with some additional privileges, which keep our own apps exclusive).

Intense internal debate

There is always the data security threat when you do something like launching a public API. And as always, there were two warring camps within Zomato, one for and the other against this move.

The logic of the camp against this move:

Logic 1: This will result in a zillion sites which will replicate our data and result in loss of traffic for us; also result in noise in the market which could be a temporary loss for our sales force. Counterlogic: Any new venture which wants to steal our data will not need our API to do it. They can use custom crawlers to do it. Heck, the labor rate in India is so low, humans crawling the website and copy pasting data will be cheaper than building a custom crawler. So we don’t have to worry about miscreants. They have always been and will always be there. Logic 2: Some of our competitors might use our API to update their own outdated data, and when they use our data, their sites will suddenly start looking better, giving us a run for our own money. Counterlogic: Most of our (serious) competitors are mature enough to not resort to such things. Anyways, to pre-empt that, we have built custom bugs in the content served to API users (yes, 0.5% of the data being served through our API is intentionally wrong), which will help us recognize and prove that this data was originally ours. Since we will have algorithmic proof to generate the data being shown on such competitor websites. Logic 3: Do we have serious developers in the ecosystem who can come up with brilliant uses of our content? Counterlogic: More than enough of them. Anyways, some of the largest users of our API will be bigger companies like Makemytrip, Mapmyindia and upcoming augmented reality startups like Mateongo. Moreover, since this blog post is a little late, I am happy to report than more than 40 developers have already requested for our test API key to develop innovative local apps :-). Logic 4: We don’t get anything out of it.. do we? Counterlogic: We get to create an eco-system of developers and apps who use our content and also help us create content for our platform. More content = happier customers. Happier customers = more customers. More customers is what everybody wants. Counterlogic won.