Tweet From The Chaff: A Smarter Social Filter

If you view your Twitter and Facebook social streams as a firehose, the filtering app Refynr fits it with a gardenhose nozzle that can be adjusted to a fine mist or a trickle. Created by the one-man start-up team Aaron Longnion, the Web and mobile app seeks to tame Twitter, allowing you to filter tweets from the people you follow based on up to 20 keywords you are interested in (and 10 you don’t want to see). While the goal is to eliminate noise and optimize your stream, the approach is more one of filtering in things of interest, rather than similar services’ filtering out of annoyances (though, of course, that’s a byproduct — and “oh, bai, @charliesheen”).

The object is to serve the most relevant information to you, and bubble up tweets of interest you might have missed. The affect is not dissimilar to what happens when you are standing in a crowded conference room where everyone is shouting and then you convene behind closed doors with a much smaller group. You can still here the rumble through the wall, and your conversations are informed by it, but the volume is turned way down.

While for most people this is probably a complementary tool, not a go-t0 Twitter and Facebook app, there are many groups looking for very specific things who may find the filters indispensable: Web programmers like Longnion and publicists come immediately to mind.

To get you started once you sign in Refynr offers to auto-generate a filter list based on the top 20 words you use most often on Twitter (but don’t be upset when Charlie Sheen shows up on it despite your public protestations).  There’s also an additional semantic option option to include any tweets that appear to be questions (containing “how,” “where,” “who” etc.).

It also suggests default options like excluding tweets containing the Fourquare check-in URL and #FF and #FollowFriday in addition to your 20 terms of inclusion and 10 of exclusion. “Although, those filters do very little work because the Inclusions do 98% of the work in first place,” Longnion says. An added option the tool offers is daily email of your filtered stream — a sort of create your own e-newsletter.

While you may at first be startled by how few tweets are in your filtered stream you are also not missing anything that the filter picks up — there is no “gap” to mind. The search function can also come in handy when you realize days (or weeks later) that you needed the information in a post. Longnion explained to AllTwitter: “Refynr starts saving all the filtered tweets and Facebook posts from your own Twitter Home Timeline and FB News Feed from as soon as you save your first Filter List. And it saves the Refyned feeds permanently for you to read later or use the Refynr Search to find something (in case you can’t remember who sent you a tweet about Product X, for example).”

Longnion is currently offering Refynr in invite-only beta. You can request a sign-up code from @refynr.

When you create a new filter list on Refynr, in addition to offering your top 20 most-used words it also tells you what 20 terms are currently most used as inclusion filters on the service. We asked Longnion what people were requisition to exclude though. The current top 5 are “job,” “iphone,” “sxsw,” “#winning” and “porn.”