SpamBayes was an open-source anti-spam filtering software that used a statistical approach based on the work of Paul Graham. Although initially developed as a tool for Microsoft Outlook, it was also offered as a plugin for other mail clients, such as a POP3 mail filter or an IMAP mail filter. In addition, an application called Hammiefilter/sb_filter could be integrated into Procmail environments.

The software worked by training the user to identify spam and legitimate mail, from which the software could analyze incoming messages and assign them a classification of "spam," "not spam," or "uncertain." Messages classified as spam or uncertain could be automatically moved to a different folder. It extracted the most significant clues from the message content to arrive at this classification and combined the probabilities to calculate a spamness score.

SpamBayes was developed in the early 2000s when spam became an increasing problem for email users. At that time, the available anti-spam filters were quite limited and often used techniques such as keyword identification or mail header verification to determine whether a message was spam. These methods could be effective in some cases, but they also resulted in many false positives and were easy for spammers to circumvent.

SpamBayes' goal was to provide a more advanced and adaptable anti-spam solution. Rather than relying on simple and easily evaded techniques, the software used a statistical approach based on the work of Paul Graham, which allowed the content of messages to be analyzed and a spamness score to be calculated based on the likelihood that the message was spam. This allowed it to adapt to spam changes and avoid many false positives that plagued other antispam filters.

In addition, SpamBayes offered an easy-to-use training interface that allowed the user to teach the software about new types of spam manually. This made the software more effective in fighting spam and more suitable for users who had specific spam detection needs.

Although it was very effective in fighting spam, the project was discontinued in 2010 and is no longer being updated or further developed. However, you can still download and use the latest version with some email clients. Although no longer viable for real-time spam protection, SpamBayes is still interesting software from a historical point of view. It may still be helpful for those who need an anti-spam filtering solution for test environments or personal use.