Ask HN: Review my Twitter app, chilitweets.com (NodeJS, MongoDB, Redis, Python)

Chilitweets is a Twitter web application I have developed over several months now. It discovers the most important links in your Twitter timeline. It basically is a tool for Twitter pro users. The basic idea of the algorithm behind Chilitweets is to aggregate all tweets in your timeline based on unshortened and normalized URLs. What you get is a personalized news feed directly generated from your Twitter account.

I’ve spend a lot of time on improving the UI of the URL list and on scaling the server backend. The server load is extremely high. The backend has to process huge amounts of tweets and URLs. They need to be unshortened and the page titles need to be loaded. For optimal results the tweets have to be aggregated over a period of several days. Which means I have to store most of the tweets for several days, too. This leads to a huge amount of data in the database. There were a lot of technical problems which needed to be solved in order to deploy this web app. I think the technical side of this project could be very interesting to you. The backend is now distibuted over several servers and I used cutting-edge technologies like nodeJS, Redis and MongoDB. The website itself is running with Python. I basically use node for all the hard HTTP work like unshortening of URLs and loading of page titles. The performance you gain is amazing especially in combination with Redis.

Since this is a one man show at the moment and I’m more a developer than a designer or marketer. I would really love to get feedback from you especially on UI improvements and on how to market this Twitter app. The biggest problem I’m facing is that the server cost per user is very high.

Thanks a lot for any feedback.

http://chilitweets.com