Closed
Description
What problem or use case are you trying to solve?
Currently OpenDevin can perform browsing on the web, but not a wider variety of web tasks. It would be good if it could do more things, such as scrolling through pages, potentially clicking on things, etc.
Do you have thoughts on the technical implementation?
Fortunately, we are in the process of incorporating BrowserGym, which should make this relatively easy. To fix this issue we can:
- Reference the implementation of the BrowseURLAction.
- Reference the set of browsing actions provided by BrowserGym.
- Implement the set of browsing actions in BrowserGym in a similar way to BrowseURLAction