today we are launching instantpulp. You will find juicy bites of real-time gossip. We are using all of feedtrace’s technology to give you the best gossip and entertainment real-time information. Hope you like it!
Feedtrace
-
2010-07-08
-
2010-05-06
Buzztrace pre alpha …
Just wanted to share a print screen of our pre alpha … two more weeks to go …
-
2010-04-30
Buzztrace (alpha peek)
@edgar and @anibal have worked hard during the last few weeks to get Buzztrace up and running. Buzztrace is a Real Time Analytics Tool for Shared Links (or passed-links). We want to help publishers understand how their links “move” in social media, how and when are they shared, how many “imprints” do they get, who are the key distributors.
We believe that passed-links and their understanding will become what the link rank and SEO was for the past 5 years. Buzztrace focus is to provide tools and intelligence to help publishers be proactive and maximize link distribution in social media.
The architecture of Buzztrace is really impressive. It works in realtime resolving short urls, aggregating them by domain, counting, comparing and plotting (See pic below). Everything is done in machine memory using a Redis DB. As we like to say: it is an oscilloscope for realtime streams. Great work guys !
We will have a beta release in 2 weeks. We’ll keep you posted …
-
2010-04-09
feedtrace.tv
A couple of weeks ago we decided to release feedtrace.tv It uses the same technology we use in feedtrace.com but it is focused in videos, to be more precise, tweets containing links to videos. We rank them and present a list of what is hot in the last hours in twitter. You can also check what video links have been shared by the people you follow. It is a different way to show you this content. Hope you like it! We are releasing today a new version with more functionality to make it much more viral.
Let us know what you think.
-
2010-03-17
what is keeping us busy
We are really busy right now. We are working on a new and improved backend. It will rock, for sure. The team is working as hard as possible, with the guidance of @anibalrojas and @edgar.
@bdo and I will be working together in NY next week. We have several meetings. We are pretty confident the meetings are going to be a success. Stay tunned for more news.
-
2010-03-02
New release: category and language preferences
The most popular links shared in twitter are typically from the US, written in English and many times about Internet and Technology. Those should account for roughly 40% to 50% of passed-links. The other 50% is long-tail. For many people they are hard to find and specially valuable. Our job is to find them and present them to the right users …
That is the quest that we started many months ago and today we take another small step in that direction. In our new version users can set preferences on languages and categories. Users that are interested in Environment will find content easier, and users in China will be presented with content written in Chinese. It is a simple explicit setting, but obviously very necessary.
In addition, in this version you should see a performance improvements on our core functions: top 50 links, hot links on a domain, popular among followings and suggested links based on your passed twitter behavior.
We hope you enjoy it ! and let us know what you think at @feedtrace or @bdo or @jordimirobruix.
-
2010-02-09
As our CEO said today on his twitter account, we are really pleased to announce that@anibal and @edgar have joined our team. They are part of the founding partners at feedtrace. They bring to the team years of experience developing web applications, and are really involved in the Ruby and NOSQL community. We are sure they will help us make feedtrace the best application possible.
-
2010-02-04
7 days later
We launched our first version 7 days ago. Besides a bug on the login session during 3 days (that we didn’t detect!), everything has gone smoothly. We have gotten some good feedback from many users, and we will incorporate their ideas little by little. Thanks.
The app is fairly stable. The response times are good in general, but in some cases the Just Following filter takes too long. We want everything to be in the 100s millisecond range, so in some filters we know we still have work to do. MongoDB has proven to be a stable DB and we are expecting performance improvements from their next release (multi core).
We expect to release a new version of Feedtrace before the end of February. It will incorporate new personalization options and improved navigation.
Now back to work …
-
2010-01-28
MongoDB hello world ;)
After several weeks migrating our relational database to Mongo DB, we have finally a version in production. Our beta testers will see little functional differences. The good thing is that we are now in our target platform, ready to make progressive improvements to our user experience. We have a long list of things to do … most of them are related to personalization of recommendations. We will work on them and release them at least monthly.
We have also good surprises coming on feedtrace.tv and a new analytics module. We have high expectations on both.
Thanks to Thimoty, Gabriel and Marielisa for the hard work. And thanks to Anibal and Edgar for their invaluable help in tuning our MongoDB database.
-
→
we are back live, check it out here
