Em@ilCRX

Commercial; $29.95

Current Version: 1.6.5 (July 6, 2007)

Hendrickson Software Components produces Em@ilCRX (EmailCRX), a desktop-based, automated email filtering software system that uses reverse DNS validation, country filtering, user defined phrases, and a challenge response system to stop spam from making it into your email inbox.

Em@ilCRX will collect all of the email from all of your POP email accounts, and - using its reverse DNS based email rating engine system, country filter, user defined phrases, and challenge response system - Em@ilCRX will validate the email, letting through only the email that is in a user managed friends list, or passes validation, or receives a valid challenge reply. Valid challenge replies will automatically add the email address(es) of the sender to the user's friends list.

The user's email will be downloaded to a user manageable database. The user retrieves cleared messages using their email client software (e.g. Apple Mail, Outlook, etc...), is able to forward clean messages to the mobile device, or from the Em@ilCRX user interface, the user can review the spam, or have it automatically deleted, or automatically reported to the sender's Internet Service Provider.

The spam being deleted from your POP server will be a huge benefit to users that also forward their email to a Blackberry, Trio, etc., because the spam will be deleted before it has a chance to be forwarded to your mobile device.

Em@ilCRX sets itself apart from other filters in the following ways:

• Filtering is DNS based (versus content and RBL-based filtering, which is more error-prone), so it is suitable for business use.
• Filtering can be done by country (e.g., filter any email originating from China, Korea, etc.)
• Challenge response is only used on messages that pass DNS validation, but the sender is not in the user's friends list.
• User defined phrases filtering is supported with regular expressions.
• Will keep both your desktop and mobile device inboxes free of spam.

Version 1.6.5 adds/changes the following:

• Changed the way the filter is trained, and added an admin item to "tune" the filter.
• Items marked as spam using the other Em@ilCRX checks are not automatically added to the filter content any longer. This approach was causing the filter database to grow unnecessarily large. New items are added ONLY when the user manually marks (or re-marks) an item as spam.
• Added a "Tune Filter" feature, that will remove obscure items (items that have been referenced an insignificant number of times) from the filter database.
• Both of these changes should improve the Bayesian type filter's performance significantly.

See the 1.6.5-related discussion thread for more information.