Everyone deserves to feel safe online. But, for better or worse, everything is out there on the internet, including harmful content. Warn helps users take control over their well-being by filtering out content that could reawaken trauma.
Warn is a browser extension that helps users pick what they want to see on the internet. It currently supports six common triggers that often appear in media: adult content, hate, gambling, substance abuse, visually disturbing content, and violence. When a user loads a webpage, Warn checks it for sensitive media, then blocks it out if it's present, without affecting the rest of the page's content. Currently, Warn blocks triggering photos, audio, and video.
Warn is a Chrome extension, written in pure JavaScript. It is supported by a RESTful API writtent in Python Flask and powered by the AssemblyAI API and AWS Rekognition.
Chrome extensions are quite challenging to build compared to a typical web app-- we ran into a lot of deprecated code, blocked functions, and especially CORS issues.
We're really proud of how we integrated AssemblyAI into this project. Being able to filter dynamic media, not just images, makes Warn far more powerful and useful to users.
We'd like to improve our recognition and add a lot more customization to the extension. Every person is unique, and so we'd like to give everyone options for what they see and what they don't.