Signal BioLogic
Signal BioLogic uses AI to triangulate social trends with clinical data, providing real-time, evidence-based validity scores for health innovations.
YouTube Video
Project Description
Signal BioLogic is a high-fidelity OSINT (Open Source Intelligence) command center that transforms the way we verify health trends. Moving beyond simple text-based chat, our application utilizes a sophisticated A2UI (Agent-to-UI) architecture to deliver a functional, interactive dashboard that responds dynamically to agent research.
Technical Execution:
The project is powered by a Java 21 MCP (Model Context Protocol) server that handles complex, secure SSL-encrypted data fetching from PubMed, Reddit, and Hacker News. By bridging this Java-based “Intelligence Spine” with a Node.js CopilotKit relay and Gemini 1.5 Pro, the agent doesn’t just provide a summary—it constructs a visual dossier.
Generative Experience:
When a user initiates research, the agent orchestrates a multi-step triangulation process. It then invokes specialized React components to render:
- An Executive Briefing hero section.
- A Momentum Graph (Recharts) comparing community hype vs. discussion.
- A Confidence Engine that visually scales based on scientific validity.
- A Source Archive table with staggered Framer Motion animations.
By using CopilotKit for state synchronization and MCP for tool-use, we have created an interface that “squishes” and adapts as the agent works, ensuring the user is always presented with an actionable intelligence terminal rather than a stagnant chat log.
Prior Work
Attended this training https://www.deeplearning.ai/alpha/short-courses/build-interactive-agents-with-generative-ui