CNET también está disponible en español.

Ir a español

Don't show this again


FaveBot intelligently hunts down content you're into

Set up feeds of topics you're into and FaveBot will track them down for you.

FaveBot is a service that keeps an eye on whatever keywords you give it to pull up related items from the Web. If you're familiar with Google Alerts, the idea is similar. In Favebot's case, you can take any keyword or set of keywords and apply it to the types of content you're looking to keep an eye on, be it photos, videos, blog posts, or podcasts. There are nine categories in all, and the system is designed to serve it up like a river of news with the most recent items appearing on the top.

What's neat about FaveBot is that it uses location as part of the filtering. For example, if you live in Texas and earmark a word like Web 2.0 and opt into the events category, you'll get the heads up on when the service finds local happenings that match up with that category (e.g. SXSW). By adding more types of items on your tracking list, you can increase your chances of seeing them in the main news stream.

Each set of keywords can be dialed in by what type of content you're looking for. Each set also has its own RSS feed.

Besides the river of information that's all mixed up, each keyword you're tracking has its own RSS feed. These can simply be subscribed to in Google Reader or whatever RSS catcher you happen to be using if you feel like skipping the site entirely. Each item also has its own permalink so you can share it with friends in the same place as your other items from that feed, although my guess is that people will simply link back to where the content was originally found.

Speaking of which, the sources are from all over, but there's a blacklist curated by humans. This is an attempt to keep some of the spammy sources from pervading your news stream. The service also recommends you combine three or more keywords together (unlike I did in the screenshot above) to avoid getting irrelevant results. Putting in "Webware" as a keyword in all of the categories brought in a good grouping of related content, although as warned, some items ended up being less focused.

I've looked at a few other services that do similar things but just for news (see and Tiinker). There's also Persai, a machine learning search tool put together by the guys from Uncov that does a more advanced version of this by putting several keywords together into packages for you and learning what content to serve you based on what you like and dislike. In either case the difference between human recommended content and machine is starkly different, and worth giving a go if you're on the hunt for new content to enjoy.