Empowering Users with Custom Spam Filters: A Cross-Platform Solution for Personalized Message Control

Have you ever wanted to take more control over your spam messages? Maybe sometimes you have noticed that your existing filters let slip a couple of spam messages, or worse, that some important messages mistakenly get flagged as spam? With (insert name here I can’t remember), this problem goes away. By allowing the user to create fully custom filters with keywords, the power is in the user’s hands to determine which messages get flagged for spam or not.

Challenges:

One of the main challenges was getting this app ready for both Apple and Android products, as Android apps are not supported natively on Apple devices and vice versa, so essentially writing two completely separate apps was a big challenge.

Another challenge, specifically on the Android side, was gaining access to incoming SMS messages and their content to analyse and filter them in the first place. Android’s proprietary SMS handler was difficult to find and use, and eventually we found an existing open source SMS app to modify according to our needs. Due to the time constraints of the Hackathon, we decided this was the best course of action, modifying an app that already managed to bypass the Android SMS handler worked much faster than spending time to do it ourselves.

Deciding on the strictness of the filters was another challenge we had to solve, for example having a filter like “Congratulations!” is a good idea, many spam messages start with that word, but there might not be an exclamation mark, or it might start with a lowercase letter. Deciding on case sensitivity, and context reading was a big challenge.

Implementation:

On the Android side, the final product allowed the user to create an unlimited amount of custom filters, with unlimited character lengths, and could contain any ASCII characters. We encouraged the user in the app to make the filters more granular, like single words or phrases that are commonly used, and acknowledged the risk that including certain words might filter out real important messages. However, using phrases mitigates that risk. For example, only using the word “Winner” as a filter, might filter out a real message letting someone know they have won something, or a message from a friend talking about a sport. Using a phrase like “You are the lucky winner” limits that risk,  as this is a common phrase in spam messages.

Within the messaging app the filters were easily accessible, and the user could enable, disable, add, edit or delete filters as necessary.

Results:

The app effectively and consistently filtered out messages according to the provided filters on both the Android and Apple versions of the app. The app integrated seamlessly in the messaging flow, was not intrusive and could even work adjacent to existing proprietary filters on Apple and Android. The app reached its goal, by allowing users to take control over their spam filters.