safechat

Inspiration

Develop a tool to detect toxic messages to create a more enjoyable environment on the internet.

What it does

Creates a less-toxic and more enjoyable and inclusive environment on the internet.

How we built it

Natural language processing (google cloud API), angular (for front-end), microservice architecture, web-sockets (for fast real-time communication).

Challenges we ran into

Finding appropriate datasets to train, processing/training of the datasets (time-constraints), server configuration, handling of client-server communication requests.

Accomplishments that we're proud of

Above 95.776% accuracy!

What we learned

Processing large datasets. Building full-set web applications. Handling diverse technology stack. Google cloud APi.

What's next for safechat

1. Faster data processing.
2. Chrome extension?