Gorantula – multi-agent AI research platform with parallel web crawlers (github.com)

Gorantula, an innovative multi-agent AI research platform, has been unveiled, featuring advanced web crawling capabilities to streamline investigative research. This AI-powered system deploys eight concurrent scraping workers, termed "Legs," to simultaneously gather information from multiple sources. It boasts a unique Cross-Case Synthesis Engine, utilizing a persistent Inverted Entity Index for historical analysis and real-time notifications about relevant past findings. A notable inclusion is the RAG Vault Chatbot, enabling users to interrogate archived investigations, ensuring contextual precision through strict evidence constraints. Significantly, Gorantula enhances user interaction with its 3D WebGL Spider View and an interactive Detective Board, visualizing gathered intelligence dynamically. This platform employs a sophisticated multi-agent persona analysis, running various specialized AI personas in parallel to extract diverse insights, including anomaly detection and causal relationship mapping. Moreover, it features robust fallback mechanisms for seamless operation amidst API failures, provider-agnostic integration with leading AI models, and a detailed discovery review process. Gorantula represents a significant leap for the AI/ML community, potentially transforming the way researchers interact with and visualize complex data, improving the speed and accuracy of investigative workflows.