What is Orca?
Orca is an agent-based browser developed by researchers at the University of California, San Diego (UCSD). It treats webpages as malleable materials and the browser as a flexible space, enabling large-scale, user-driven browsing with AI assistance. Users can view multiple webpages side by side on an infinite canvas, while AI agents handle repetitive tasks such as clicking buttons or filling out forms. Orca supports large-scale viewing, management, navigation, and information extraction, empowering users to organize webpages more flexibly, reduce the cost of managing multiple pages, stimulate exploration, and strengthen control and trust over information sources.
Key Features of Orca
-
Spatial Multi-Page Management: Arrange multiple webpages spatially on a canvas. Users can freely drag, resize, and view them side by side for easier comparison and integration.
-
AI Agent Automation: The built-in AI agent automates repetitive tasks like button clicks and form filling, reducing manual operations and improving browsing efficiency.
-
Large-Scale Information Extraction: With natural language instructions, users can quickly extract key information from multiple pages. The extracted data is displayed as unified cards above each page for fast scanning and comparison.
-
Context Expansion & Bulk Navigation: Based on the current content, Orca can suggest related pages automatically and open multiple relevant links in bulk, accelerating exploration.
-
Dynamic Workspace Customization: Users can group webpages into personalized dynamic workspaces tailored to their tasks, supporting real-time parallel viewing, flexible extraction, and dynamic reorganization of content.
Official Resources for Orca
-
Official Website: https://orca.jiang.pl/
-
arXiv Technical Paper: https://arxiv.org/pdf/2505.22831
How to Use Orca
Orca is currently in development. Interested users can visit the official website https://orca.jiang.pl/ and join the waitlist.
Application Scenarios of Orca
-
Travel Planning: Compare multiple booking websites side by side, extract key details, and quickly identify the best options.
-
Academic Research: Open multiple academic databases simultaneously, with AI agents assisting in downloading and extracting abstracts to improve efficiency.
-
News Reading: Compare reports from different news outlets and extract key points through natural language commands for real-time updates.
-
Market Research: Collect and integrate data across multiple platforms, extracting key insights to streamline research and analysis.
-
Project Management: View multiple project-related webpages in parallel, consolidate information into reports, and delegate repetitive tasks to the AI agent.