RegenReef DeSciAI

RegenReef DeSciAgent is an advanced AI-powered platform designed to support coral reef conservation and restoration. By harnessing the capabilities of large language models (LLMs), the system offers informed decision support and intervention planning for a broad range of stakeholders, including scientists, policymakers, NGOs, and local communities.

The platform aggregates and analyzes diverse data sources, including ecological monitoring data, climate projections, and socio-economic indicators, to generate actionable insights for coral reef management. (Note: While current data inputs are hypothetical for development purposes, real-world datasets will be integrated soon.)

RegenReef DeSciAgent offers several powerful features:

🌐 Natural language Q&A for accessible knowledge queries

📍 Restoration site prioritization based on multi-layered data

📝 Automated reporting for funders and regulatory agencies

📊 Scenario modeling to forecast outcomes and optimize interventions

These tools enable users to plan, monitor, and execute conservation strategies with greater speed, precision, and scalability than ever before.

🤖 Automation and Scale AI-driven tools within the platform automate time-consuming and resource-intensive tasks, such as:

Underwater video analysis

Sensor data interpretation

Coral and habitat classification

This automation makes it possible to map, monitor, and assess reef ecosystems at scale. It significantly reduces labor costs and improves accuracy. It also helps in early threat detection, optimized restoration site selection, and standardized reporting, ensuring transparency and reproducibility.

🌍 Collaboration and Community-Driven Science Beyond technical tools, the platform fosters collaboration, knowledge sharing, and adaptive management across the global coral reef conservation community. By providing stakeholders with accessible, data-driven recommendations, RegenReef DeSciAgent promotes long-term reef resilience and supports the well-being of communities that depend on healthy reef systems.

🧠 Knowledge Base and Point System The current version includes a knowledge system that references and synthesizes information from top coral reef research papers, articles, manuals, and best practices in reef management and conservation. This knowledge base can be accessed through our GPT-powered agent here: 🔗 RegenReef AI Agent: https://chatgpt.com/g/g-68688d70f9288191bcaf63f961032e06-regenreef-aiagent

The platform also includes a prompt-based point system with ranking, that will evolve into a governance token model. In the near future, this system will allow for:

🧪 Community-driven research funding

🧬 Support for independent researchers

🌱 Incentives for real-world ecological impact

This design aligns with DeSci values, focusing on decentralization, transparency, community governance, and open access to knowledge. RegenReef DeSciAgent ultimately aims to support nature-based solutions and build a regenerative, stakeholder-powered response to the global coral reef crisis.

this is a prototype, it is not ready at all :D

RegenReef DeSciAgent was built with a focus on modularity, user accessibility, and integration of cutting-edge AI tools for ecological impact. Here's a breakdown of the stack and approach:

Frontend & Logic: We primarily use TypeScript to ensure strong typing and maintainability across the codebase. It powers both frontend interfaces and backend logic.

AI & Reasoning Engine: We integrated LLMs (Large Language Models) to act as decision-support agents. These Deep Agents analyze context from coral reef data and deliver structured responses. These agents are trained and fine-tuned (in-context) on a corpus of coral reef conservation best practices, scientific literature, and ecological data modeling techniques.

Authentication & Onboarding: For a secure and smooth user experience, we integrated Privy.io in the onboarding process. This allows users to connect via:

Ledger hardware wallets

WalletConnect

Rainbow Wallet

Email-based login

Multi-Step Verification Mechanism: We added a double-verification layer using Flow Wallet on Flow Testnet. This serves two purposes:

Verifies identity and activity of users contributing to the knowledge base

Assigns and tracks participation using a custom point system embedded in the user dashboard, that in a future could be quite tokenized for proper incentives.

📚 Knowledge Dashboard The platform includes a dashboard that connects users to curated topics related to coral reef conservation. These are drawn from high-quality sources including peer-reviewed research, technical handbooks, and reef management guides.

Each interaction with selected topics or challenges contributes to a scoring mechanism that will later evolve into a governance or participation token model—helping distribute real value to contributors.

🛠️ Hacky and Notable Integrations

We built a "prompt-to-score" engine, where interactions with the LLMs on coral reef topics generate internal metrics.

This prototype mechanism can: Track how deep or accurate a question is Trigger engagement milestones

Recommend next topics automatically based on learning behavior

We also implemented real-time LLM embeddings caching to reduce redundant computation and API calls. This was essential for sustainability, especially when working with ocean conservation content that doesn't change frequently.

🤝 Partner Tech Benefits

Privy.io: made onboarding smoother for non-crypto-native users (scientists, conservationists, etc.) by offering flexible sign-in methods.

Flow Blockchain: was chosen for its lightweight verification architecture, ideal for scaling reputation and scoring systems without heavy gas fees.

Ledger Wallet was used for security purposes as part of the double verification process and Flow Wallet integration.

OpenAI’s API (GPT-4) was the backbone of our AI-agent for contextual analysis and user interaction.

We use Netlify because it offers a user-friendly dashboard to control everything related to our hosted websites, including builds, deployments, custom domains, serverless functions, and team settings. It allows us to fuse static website hosting with frontend frameworks, domain management, and powerful backend features like functions and webhooks—all in one seamless platform.