Intelligent Browser Agents

Discipline: CS

Advisor: Liqiang Wang

Intelligent Browser Agents aims to solve the problem of users repeatedly performing time-consuming tasks on web browsers, which can lead to fatigue and input errors over time. This project utilizes AI agents to perform actions on browsers according to a user's prompt to plan, execute, and verify actions. It was built for students, business professionals, and anyone who completes repetitive tasks on the internet. The system uses five AI-driven agents to reason through tasks before/after taking actions. A user submits a request through the frontend interface, where the orchestration agent creates a plan to accomplish the goal. The execution agent performs browser actions using tools to capture page information and interact with websites. Verification and fallback agents ensure the task was completed correctly. Our project improves web-automation through natural-language interaction and a multi-agent architecture. Example use cases include scheduling appointments, filling out forms, planning trips, applying for jobs, and conducting research. The system also provides real-time task monitoring and is built on a scalable architecture using FastAPI, PostgreSQL, and AWS. Intelligent Browser Agents provides a new way for users to interact with computers while reducing human error and increasing productivity.

Back to Showcase