getFilter 2

My girlfriend is in advertising. But she kept being disturbed by unwanted call and junk messages for years. To clarify that promotion calls and text messages are not real advertising. So we decided to make getFilter.

● How we built
◎ Machine learning:
A Core ML 2 model trained with more than 300k text messages in 3 different languages has resulted an AI with more than 96% accuracy that can filter unwanted text messages. We support English, Chinese, Persian.

◎Safe and anonymous Crowdsourcing:
Every-time a custom rule has been created by you or a message/call that you report with the new iOS12 feature anonymously that sender receives a rate on getFilter servers. Twice a day your app background refresh receives updates for all of the reported numbers.

◎Custom rules:
You can custom rules by creating keywords or phone numbers to block/allow specific text messages and calls.

● How it works
When unwanted calling or junk messages come, getFilter applies Machine Learning to check first. Then getFilter goes through crowdsourcing list to strengthen defense. At last it matches calls or messages based the rules you created. After 3 layers filtering, here we block everything you don’t want.

● Filtering criteria
For text message, every message goes through the enabled methods enabled by you. If all methods are active each message goes through all these methods to get either blocked or not.
1. Your rules, If any text message or sender matches your custom rules. it will get applied to that message.
2. Crowdsourcing, If any sender matches any of the numbers in crowdsourcing with the CS rule condition u have enabled will get blocked or allowed.
3. Machine learning, if the two previous methods have applied no conditions we use machine learning. Before each machine learning use we remove numbers, links and special characters and first online with more powerful AI and if there are no internet connections or our server fails we use a smaller AI offline.

※Privacy Policy: What information we collect?※
getFilter won’t collect any personal information about your or do profiling. Show More