A new realtime search engine called Wowd is launching publicly today at the Web 2.0 Summit in San Francisco. Founder and CTO Boris Agapiev gave the first public demonstration of Wowd at our Realtime Stream Crunchup last July (see video below).
Wowd takes a very ambitious approach to search in that it is a peer-to-peer search engine. Users download a peer-to-per client and the index exists not on any central cluster of servers, but across all the user’s machines. Agapiev, who formerly founded the vertical search engine Vast, set out to conquer scaling issues in a new way and settled on a P2P approach.
What makes Wowd a realtime search engine is that ranks sites based on how often and recently the Wowd community has visited that site. On Wowd, you vote with your mouse, so to speak.
The search engine uses other more traditional ranking algorithms as well, but its main point of differentiation is the realtime clickstream data it gets from the peer-to-peer clients which it wants users to download. In theory, this will provide Wowd with your complete attention history if you allow it—every site you go to, not just the ones you click to from the search engine (in this regard, it is similar to what the Google toolbar records if you have the Web History feature enabled).
As with any attention recorder, Wowd offers a full range of privacy settings so that you can share only what you wish, and it is all anonymized and Wowd doesn’t even know your IP address anyway. You can always see and search your entire Web history, which is helpful when you are trying to remember that obscure furniture site you saw last month. Nevertheless, this will be a big barrier for many consumers who might not feel comfortable sharing their surfing habits with an unknown startup. Simply asking people to download a client will be a barrier to adoption.
But for those who do, they will be presented with a slightly different search experience. Results are ranked based on which sites are most popular with the Wowd community. Sites that have been visited recently get a stronger weighting. You can also switch to see the freshest results. The quality of the results depend on the Wowd community’s finding and visiting the best sites, but it is all based on passive activity.
Wowd’s results supposedly won’t be as susceptible to SEO spam as other search engines. But if it becomes popular enough I’m sure spammers will try to game it by simply getting lots of people to download Wowd and visit their own sites continuously. Wowd obviously tries to monitor this type of behavior and weeds those clicks out from results.
I’ve been testing Wowd for a coupel months, and the results are decent already with only a few thousand beta users. Now that Wowd is open to the public, the real test will begin. Realtime search is hot right now, with tons of startups (Collecta, OneRiot, Topsy) and as well as bigger companies like Twitter, Facebook, and Google going after it.
Below is a video of CTO Boris Agapiev’s demo at our Crunchup in July. It starts at the fourth dot along the timeline at around 29 minutes.
That site hurts my eyes. And I think I can taste this mornings breakfast.
Too complicated for too little payoff.
I think I will stick to http:/www.yauba.com for my real time search needs
That site looks pretty cool…it was the only one to find info on a level 3 outage that happened this morning!! It showed up in the real-time search
Yes Yauba rocks.
Yeah.
I will not download anything to help another site.
Also, the 1-20 things arn’t necessary.
This rocks
Congratulations on the launch! I’ve been using Wowd for a few weeks now and it’s very cool. As I’ve told the Wowd team, the P2P search concept is very compelling.
By the way, I’m not seeing the video.
video works fine for me. Anyone else having trouble seeing it?
What video?
I’m not a coder so I may be wrong about this, but why couldn’t they just track user visitation data through a cookie? Isn’t that how Google works?
Google tracks which sites you visit after you do a search. I’m guessing that the idea here is to track *all* the sites you visit, which is much more powerful.
Good idea from an engineering standpoint. Obviously scalability becomes less of an issue.
Too bad about having to install desktop software. I’m honestly not sure if the value proposition of realtime search is high enough for me to go to the extreme of download, install then trade away my computing resources.
Wow. Sounds just like Faroo (http://Faroo.com) that’s been around for at least a year now http://bit.ly/1BcL2B
Cool concept. Anybody who helps keeping Google honest is my friend! These are smart guys – and it works.
I think it’s cool that they’ll only get better as more people join the network. The peer-to-peer model is extremely interesting and could gain traction if they deliver quality and timely results. I can’t wait to see how Wowd’s doing in 6 months.
Thanks to all for comments, we are very excited, there is so much more to be said on search
Regarding tracking through cookies, it is one way but does have drawbacks, certainly scalability (as one of the previous posters mentions), also latency.
Twitter is really a great example of what we consider an “explicit” attention frontier, where users vote with their (explicit) actions (tweets with links). It has certainly been quite amazing to see what has been already demonstrated using this data in real-time search.
What we do is more of “implicit” attention frontier, where the simple act on clicking on a link is considered a vote. This way it is very easy on users, and there are many such signals.
We are really excited about great things coming up in real-time, distributed search & discovery
See Steve Jurvetson (an investor) discussing the concept here: http://www.yout...PDbOUH1I#t=0m6s
This little talked helped me envision the concept a little better, not to mention Steve’s pretty efficient when he speaks. It’s interesting to see a concept that could give Google competition at scale. Do you guys have patents on the technology or is it more of a “whoever reaches a critical mass with users first wins” scenario?
Using peer-to-peer is only as good as 1.) bandwidth and 2.) firewall rules. Both dimish the ‘real-time’ claims of this. Additionally, I don’t trust any p2p app. I used to work at a company that tried ’secure p2p content distribution’ – and we could populate whatever data we wanted on any client / cache as we called them. So my trust level of p2p = bitbucket. I don’t see techy types using anything other than google for a long time
John Battelle is currently talking about wowd on #w2s
Boro Cograts!
John Battelle is currently talking about wowd on #w2s
Boro Congrats!
Interesting, I’ll have to take a look. It seems to use crowdsourcing techniques.
Pretty interesting ideal but using the external application will limit its use in my opinion. If they could create a firefox add-on with the same functionality I think it would be more widely used.
Sucks, real bad.
I’d take you so much more seriously if you could spell “couple.” The fact that you can’t seriously raises questions for me about your journalistic ability as a whole.
Id have to say they are off to a good start!
RT
http://www.anonymous.ua.tc
Wowd is great…using app. on Mac and FF week ago still work fine and fast, without bugs, memory or network resource threats.
I’m not suspicious about Wowd p2p and distributed clouds based technology, and hope marketing team work hard to present this new search engine.
Good luck guys!
no one is going to use this, another program running in the background logging what I do? No thanks.
Google, Yahoo…do same, putting cookie, collect you history and even IP. Normally you can switch off this default setting on Google or browser. So, you can do it on Wowd too.
Classic SE collect data on central server (danger for you privacy), while Wowd not save this data on any central place. That’s benefit to you privacy.
But the potential of a desktop app to hurt the user that installed it is much larger than what you can ever achieve with cookies etc.
Maybe you are right. That’s a question of trusting or not trusting company. Let me joke… anyway, best way you protect yourself…escape from Web.
This thing is DOA just because of having to download and run an application alone. Why make search more complicated than it needs to be?
The benefits of peer-to-peer search engines is that they can save money to the founders, but they can’t provide better search results than Google, Bing and others.