We were inspired by hotkeys and aimbots for games and wondered how other actions like mouse movements could be made more efficient. Web surfing is mouse intensive, especially for pages with an abundance of links. When surfing the web for an extended period of time, all the wrist movement could add up, resulting in muscle strain. So, we decided to ideate on a tool that would allow users to browse the web on cruise control.
What it does
Based on the user's mouse movements, Cruiser Cursor will try to find the most relevant clickable object in that direction. The object will be highlighted for the user and they may click wherever they are to simulate a click on that object.
How we built it
Once activated, the extension takes snapshots of the user's cursor at fixed intervals. Using this, over a small window, a vector is created for the direction of the cursor movement. The script goes through the website identifying all possible clickable objects such as buttons, images or links.
Using information from the direction, the algorithm finds the most relevant clickable object within a fixed field of view from the cursor and animates that object. If the user likes the prediction, they can simply click at their current cursor position, taking them to the next page and expediting the browsing process!
Challenges we ran into
- Some Chrome extension API was deprecated
- Not all pages built the same, so HTML/CSS part did not work perfectly on all components of all pages
- Difficult to communicate as we were doing the hackathon virtually
- Integration with servers like Firebase proved difficult owing to security constraints
Accomplishments that we're proud of
- Getting a MVP out
- Learning how to make a Chrome extension
- Creating a product we find useful
What we learned
- Chrome extension API
- A more in-depth understanding on front-end languages like HTML/CSS
- Math proofs to implement algorithms to detect and predict most likely clickable elements in web pages.
What's next for Cruiser Cursor
We want to incorporate a consensus mechanism which tracks links users click on every webpage. We will then use a balance between proximity and popularity to determine our predictions, making Cruiser Control a lot better!