I started writing Conflux on August 30th; came up with the idea a couple days before. Was unable to find anything quite matching my idea, so I decided I had to implement it (and had the time to do so, since I have not been working a regular job :). Having sunk a lot of time into developing both the idea and the software, the first public release is now available from the Marble Horse Free Software Group site and from its primary home at SourceForge.
Primary
distribution page for Conflux
Alternate distribution page for Conflux
Advogato project page
Conflux provides a number of useful features, both for the individual and for workgroups; lots of potential applications for this code. The basic notions are this: a perl backend (which may be run as a daemon, via cron or standalone), the collector, which performs searches on your behalf and posts the resulting URLs to a central database; a PHP3-based frontend which allows you to submit searches (to be performed now or later), view results and bundle together useful searches and search results. A number of more advanced features have already been implemented, including a family tree model for tracking search history.
A number of innovative (I think) features are planned for Conflux. The basic framework and core functionality is done, so I decided to release the first public version. Conflux currently works with MySQL (via perl DBI and php3 w/MySQL support); other database flavors will be supported in future versions (starting with postgres and DBM). Hopefully others will find this to be a useful tool :) Conflux 0.11 has been released under the GNU GPL v2 or above. Though there is enough documentation to get Conflux set up (and the user interface is mostly self-explanatory), I've been working for the last couple of days on a high- quality manual which I'll eventually release under the GNU FDL 1.1 or above. Full manual should be included with the next public release.