Allth.at was one of the 100 top alternative search engines in Charles Knight's February list. It's a search agent that allows you to define a search topic, then refine it with the use of filters, and finally subscribe to it via email or RSS. It's similar in many respects to PubSub, the now defunct 'future search' engine that I was a fan of. I got the chance to catch up with Allth.at founder Morgan Snyder recently, to find out more.
How Allth.at works
When you go to Allth.at, it at first looks like a normal search engine. This is a good thing, as it means the UI is simple and intuitive (something PubSub never achieved). First you input your desired search - say "web 3.0" ;-) Then you filter, either by using the search sources tabs (made up of common search sources like Google, Yahoo, eBay) or by applying graphical boolean filters, i.e. the drop-down boxes on the left. You can add content sources by clicking the 'Add' tab - for example you can add readwriteweb.com as a source (as I did for the example below). And with the drop-down boxes, you can filter words in or out - including clicking on the 'specify' link to add your own terms.
After you're done filtering, you can save your search and choose an alert if required (note: you need to register for an account first). While you can simply come back to the website every now and then and check the latest results of your "saved searches", it's easiest to subscribe to an alert. This of course automates the delivery of new search results, using two methods - subscribe to email notifications, or RSS feeds.
You can also filter your saved searches at any time - and this automatically updates the alerts (email, feed). Morgan told me that Allth.at can also filter out the duplicates.
Web 3.0 Goodness
Morgan pointed out to me that Allth.at is a good example of what Alex Iskold wrote about in his recent popular post Web 3.0: When Web Sites Become Web Services. That post was about what kind of applications will be born once many APIs are opened up (it wasn't about defining 'web 3.0', as some commenters claimed). Morgan says that Allth.at is a good example of what Alex was driving at:
"Allth.atÄôs goal is to integrate every searchable API in existence. Obviously every website does not have an API. So instead of scraping, we have taken a hierarchical approach that uses a combination of APIs, RSS and ÄúindexedÄ? results (equivalent of a site: search on google or yahoo). When a user selects a source to search, we check to see if that site has an API. If so, chances are we have integrated it and will use it. If not, weÄôll check to see if that site has a searchable RSS feed. If so, weÄôll use that. If not, then weÄôll deliver results off of the index of one of the big engines."
Allth.at is not alone in the market segment which Charles Knight has called "Continuous Search Engines". In our February Alt Search Engine list, we noted Swamii, which continually search newspapers, TV, the web (and more) to let you know when new, relevant and cool stuff arrives. There is also Searchbots.net, which gives you your own personal search robot - to continuously search the Internet, trying to find all the best websites it can, on your behalf.
You could also consider Google Alerts in this category, since they can be set to "continuous" alerts also.
I asked Morgan what is the difference between Allth.at and the competition. He told me that usability and ease-of-use is what differentiates allth.at and noted that even Google CSE and Rollyo don't deliver custom search results 'on the fly'. Morgan told me that another key differentiator from other "custom engine providers" is that Allth.at's results "are more real time". He said:
"For example, GoogleÄôs CSE delivers results off of their index. Rollyo I believe is powered by Yahoo. Because we use APIs, RSS and then index, Allth.at returns the most current results first."
Morgan explained how this works, using my example of searching for the term "web 3.0". If you use Googles CSE and select readwriteweb.com as a source, it would be the equivalent of doing a 'site: search' on Google. If you use Rollyo, and select readwriteweb.com as a source, itÄôs the equivalent of doing a 'site: search' on Yahoo. Depending on when Google or Yahoo last updated their index, they may or may not pick up the latest results (e.g. Alex's article). With Allth.at, it first searches the RSS feed for readwriteweb.com (because this site doesnÄôt have an API), which would therefore pick up Alex's web 3.0 article. Allth.at delivers indexed results on top of that too. Here is an illustration of this exact scenario:
Overall, Allth.at is a very easy to use search agent and outputs some nice 'topic feeds' for users to subscribe to in their RSS Readers. An easy-to-use, yet sophisticated and reliable, 'topic search' service is something I've been looking for a long time - and Allth.at looks very promising in that regard.
Today Allth.at takes the covers off their relatively stealth company (well it was stealth until Charles discovered it!). It's a bootstrapped little startup, with just 3 people. They took the splash page down just under a month ago and today is the launch of Allth.at's "live beta". Color me impressed, but check it out for yourself and let us know what you think.
Note: thanks Charles Knight for his help in writing this article.