The Ageless Odyssey: A Longevity R&D Decision Engine

Live Demo: https://neuroidss.github.io/ageless-odyssey/

"The goal is not to reverse aging, but to correct flawed development through superior engineering. A complete understanding of any system is the true path to its mastery." - The Engineering Paradigm

This project is a simulator for a new paradigm in longevity research. It shifts the focus from "anti-aging" to a more rigorous, systems-engineering approach. As highlighted by experts, longevity research faces a core "chicken-and-egg" problem: we need sensitive biomarkers to validate new interventions, but we need effective interventions to discover which biomarkers are truly meaningful. This is compounded by a low signal-to-noise ratio, making it difficult to distinguish real progress from statistical noise.

This application is an AI-powered decision engine designed to tackle this challenge. It frames the journey as a strategic allocation of resources to answer a critical question: "How do you most effectively invest in our engineering understanding of life itself to escape the current research plateau?"

You don't just research aging; you guide a team of AI agents to identify high-impact research trends, simulate funding R&D pipelines, and upgrade your own biological substrate. Your actions serve as a benchmark for a new form of human-AI co-evolution, with the ultimate goal of transcending biological limits through complete comprehension.

---

Core Engine: The Longevity R&D Decision Engine

The application is fundamentally a prototype for an agent-driven system to prioritize spending and investment in the science of systems engineering, as applied to biology and beyond.

What's Implemented

• Temporal Analysis Lobe: A direct implementation of the core research methodology. This view allows comparing two "snapshots" of the research space (e.g., at time t1 and t2) to visually identify new information clusters, providing a clear way to see how a field is evolving.

• Investment & Upgrade Portfolio: The central feature is the Engineering Portfolio, a dashboard for allocating symbolic capital. It's split to directly address the "chicken-and-egg" problem:

  1. System Upgrades (The "Shopping Cart"): For purchasing market-ready therapies (TRL 9) to improve the current biological system's performance and provide a stable baseline.
  2. R&D Investment Portfolio: For allocating capital towards promising, low-TRL research to discover the next generation of interventions and the biomarkers needed to measure them.

• AI-Powered Trend Analysis for Deal Flow: The TrendSpotter agent acts as a scientific analyst, identifying emerging research trends and scoring them on Novelty, Velocity, and Impact. This provides a scientific rationale for making investment decisions in the R&D Portfolio.

• Federated Data Collection: A multi-service parser (searchService.ts) gathers information in real-time from PubMed, bioRxiv, Google Patents, and the OpenGenes database. This forms the evidence base for all agent analysis.

• Agent-Driven R&D Simulation: The InterventionMarketplace simulates a Technology Readiness Level (TRL) pipeline. R&D stages are framed around core longevity challenges, such as "Biomarker Discovery" and "Preclinical Validation," directly reflecting the concepts from the hackathon lectures.

• Dynamic Knowledge Graph: For each research query, the system generates a localized knowledge graph, visualizing the relationships between concepts (genes, compounds, processes). The "Temporal Snapshot" feature allows viewing the evolution of this graph over time.

What's Conceptual or Not Yet Implemented

• Real Financial Integration: The "investment" mechanic uses a symbolic Capital resource. The system is not connected to real-world financial data or venture capital models.

• Global, Persistent Ontology: The app's knowledge graph is dynamic and query-scoped. It does not yet build towards a single, persistent, and curated global ontology.

• True Embedding & Clustering: The "Temporal Analysis Lobe" visually represents the concept of comparing information clusters over time. It does not yet perform true vector embedding and t-SNE/UMAP clustering on the backend.

---

Relation to Hackathon Tasks

The app's architecture directly engages with the hackathon's core challenges.

Task: Longevity Knowledge Graph

• Status: Strongly Supported.
• Implementation: The app's KnowledgeNavigator agent builds a dynamic, interactive graph for every query. The "Temporal Snapshot" control allows users to see how this graph evolves with each new piece of information, turning it into a living map.

Task: AI Agents for Extracting Bioactivity from Patents

• Status: Partially Supported.
• Implementation: The system includes a CompoundAnalyst agent and integrates Google Patents search. The R&D pipeline in the marketplace includes a "High-Throughput Screening" stage that uses this agent, directly connecting the task to the app's core mechanics.
• Gap: The implementation is for ad-hoc analysis to advance a research goal, not a systematic pipeline for building a large, standalone dataset.

Task: OpenGenes AI Benchmark

• Status: Partially Supported.
• Implementation: The GeneAnalyst agent uses the OpenGenes API and is prompted to extract structured data (organism, lifespan effect, intervention) directly inspired by the benchmark's requirements for high-quality, nuanced data.
• Gap: The app uses the data in a manner consistent with the benchmark's goals but does not evaluate an LLM's performance against the OpenGenes ground truth.

---

Supporting Features

• The Odyssey Framework: Progressing through Realms (from Mortal Shell to Stellar Metamorph) provides a long-term goal and narrative structure.
• Personal Trajectory Simulation: The biomarker simulation provides a feedback mechanism to make "System Upgrade" choices feel more personal.
• Achievements & Quests: These provide short-term goals that guide the user through the app's capabilities.

---

How to Run Locally

Prerequisites: Node.js

1. Install dependencies:

   npm install

2. Set up your environment (Optional, for Google AI Models): If you wish to use Google AI's models, create a file named .env.local in the project's root directory. Inside this file, add your Gemini API key:

   GEMINI_API_KEY=YOUR_API_KEY_HERE
   

   For local models (Ollama, Hugging Face), this step is not required.

3. Run the development server:

   npm run dev

This will start the development server, and you can view the application in your browser at the local URL provided in your terminal (usually http://localhost:5173 or similar).