denkwerk TechTalk

mosparo, the modern open source spam protection

Bot detection and filtering with mosparo

However, mosparo has two practical functions for recognizing bots, i.e. automated spam postings: Firstly, there is a honeypot field that is not visible to human users, but is visible to bots: if it is filled in, it is clear that an automated system was at work here and the request is almost certainly spam. On the other hand, mosparo also checks the time users need to complete the form: If the form is filled out and sent too quickly, there is also a lot to be said for a spam entry.

According to Sascha, another practical feature of mosparo is that the tool works with a so-called ruleset: This is a JSON file that contains the spam rules created by the server operator. This ruleset is not only transferable, but can even be transferred from another server via URL – ideal for offering a central anti-spam file on denkwerk servers, for example.

Our conclusion

Of course, mosparo is not free from criticism either: the detection quality of the software stands and falls with the quality of the ruleset. Sascha cited an example here in which a user was repeatedly filtered because a part of her name was classified as spam. Accordingly, when in doubt, the ruleset must be gradually adapted to the actual needs of a website or online service.

Conversely, “annoying” contact form requests that are not classic spam - such as contact requests from influencers – can also be filtered very easily in this way. However, rulesets still have to be defined by the users themselves; mosparo does not (yet) have a ready-made list for the German-speaking world: “Unfortunately, German-language spam words still have to be manually collected and entered,” says Sascha.

If you would like to find out more about denkwerk and our projects, simply subscribe to our newsletter: