When Sam, Andres and I started working on Octopart we were attracted to the idea of building everything ourselves, from a Python web framework to our own rack mounted cluster of machines.  In the ensuing years we have learned to work more efficiently by relying on countless cloud services such as Amazon EC2 and GitHub.  Today we are happy to make our own contribution to the world of software as a service with the beta launch of ThriftDB, a cloud service that allows anyone to quickly build a fast and reliable search index.

One of the largest challenges we faced with Octopart was building a search engine to help our users find parts.  In the process we noticed that search seems to be the ugly stepchild of engineering problems - it often goes ignored but has the potential to significantly improve the user experience on countless websites.  Our goal with ThriftDB is to use lessons learned from Octopart to take away all of the pain points of enabling search.

We've indexed Hacker News and built a search interface into its data to demonstrate the capabilities of the ThriftDB API.  You can read more about it here:

http://blog.hnsearch.com

We have opened up a test interface that anyone can use without an account to explore the API.  We are all really excited to release ThriftDB into the wild today and get your feedback!

Andres, Sam and Harish