An Ai DeFi assistant that helps users navigate decentralised finance through personalised strategy recommendations and investments controlled by a personalised Ai-Assistant.

Deployed Project

Project Repo

Video

• Web3 Wallet Integration – Connect via MetaMask for secure authentication and transaction execution
• Personalised Investment Profiles – Based on user’s goals, risk tolerance, investment amount, time horizon, and preferences
• AI-Generated DeFi Strategies – Smart recommendations using Google Gemini LLM and real-time protocol data
• Interactive Dashboard – Monitor strategies, track performance, visualise risk and returns
• Conversational AI Assistant – Get answers, portfolio updates, or strategy recommendations via natural language
• Smart Contract Execution – Direct integration with Ethereum-based DeFi protocols via MetaMask
• Adaptive Portfolio Management – Strategies update dynamically based on user changes or market trends

Our system generates personalised DeFi investment strategies using live data and AI, with a focus on Ethereum-based lending and staking opportunities.

• User Profile

  • Risk tolerance
  • Investment goals
  • Amount & currency
  • Time horizon
  • Experience level
  • Preferred DeFi activities

• Market Data (via DeFi Llama)

  • Protocols on Ethereum only
  • Status: Active
  • TVL > $1M
  • Categories: Lending, Staking
  • Matches user’s preferred activities (if specified)

1. Filter & Rank Protocols

   • Filter DeFi Llama data:
     • Ethereum protocols
     • Active status
     • TVL > $1M
     • Staking or lending category
   • Rank by:
     • Total Value Locked (TVL)
     • Reputation & historical APY
     • Fit with user risk/goals
     • Compatibility with user experience

2. Prompt AI for Strategy

   • Construct prompt from user profile + ranked protocols
   • Send to Gemini 2.0 Flash model
   • AI returns structured JSON:
     • Protocol name
     • Activity (staking, lending)
     • Token
     • Allocation %
     • Expected APY
     • Estimated return
     • Risk level
     • Justification

3. Evaluate & Present

   • Ensure:
     • Portfolio allocations sum to 100%
     • APY-to-risk is favourable
     • Diversification and goal alignment
   • Display as cards in dashboard with charts and execution buttons

1. Connect Wallet

   • Authenticate with MetaMask wallet

2. Create Investment Profile

   • Complete questionnaire:
     • Risk tolerance
     • Goals
     • Preferred activities
     • Experience level
     • Investment amount and currency
     • Time horizon

3. Receive AI-Powered Strategy Recommendations

   • AI generates detailed DeFi strategies
   • Includes rationale and risk analysis

4. Explore Dashboard

   • View:
     • Profile summary
     • Allocation pie chart
     • APY bar chart
     • Strategy cards
     • Performance over time

5. Execute Strategies

   • One-click execution via MetaMask
   • View transaction hash and status

6. Chat with Assistant

   • Conversational AI interface
   • Ask for clarifications or update your profile
   • Receive new strategy recommendations

Sum of all strategy returns

Shown in dashboard as:

• Total projected return
• Risk-weighted return breakdown

/
├── backend/
│   ├── main.py              # FastAPI entry
│   ├── models/              # DB schemas
│   ├── routers/             # API endpoints
│   ├── services/            # AI & DeFi services
│   └── requirements.txt     # Python deps
├── frontend/
│   ├── app/
│   │   ├── api/             # API proxy
│   │   ├── dashboard/       # Portfolio page
│   │   └── page.tsx         # Landing
│   ├── components/          # UI components
│   ├── lib/                 # Utils (Gemini, Web3, etc.)
│   └── package.json         # Node deps
├── README.md
└── .env files               # For config

Backend (.env)

SUPABASE_URL=
SUPABASE_KEY=
GEMINI_API_KEY=
DEFI_LLAMA_API_KEY=

Frontend (.env.local)

NEXT_PUBLIC_BACKEND_URL=
NEXT_PUBLIC_SUPABASE_URL=
NEXT_PUBLIC_SUPABASE_KEY=

• 🔑 Private keys are never stored – all transactions require explicit user MetaMask confirmation
• 🧾 Transaction Transparency – view full on-chain data before confirming
• 🔒 Row-level security – enforced in Supabase for multi-user data safety
• 📡 Secure connections – Web3 communications via encrypted RPC providers

• id: UUID
• wallet_address: string
• email: string?
• created_at
• last_login

• id: UUID
• user_id: UUID
• risk_tolerance: string
• investment_amount: float
• investment_currency: string
• experience_level: string
• investment_goals: string[]
• preferred_activities: string[]
• investment_horizon: string

• id: UUID
• user_id: UUID
• strategy_data: JSON
• is_active: boolean
• created_at
• updated_at

• id: UUID
• user_id: UUID
• asset: string
• amount: float

• id: UUID
• user_id: UUID
• transaction_hash: string
• blockchain: string
• protocol: string
• activity: string
• amount: float
• token: string
• status: string
• timestamp

• Node.js 16+
• Python 3.8+
• MetaMask extension
• Supabase project

# Navigate to backend directory
cd backend

# Create and activate virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# Set up environment variables
cp .env.example .env
# Edit .env with your API keys and configuration

# Start the server
uvicorn main:app --reload

# Navigate to frontend directory
cd frontend

# Install dependencies
npm install

# Set up environment variables
cp .env.example .env.local
# Edit .env.local with your API keys and configuration

# Start the development server
npm run dev

• After Gemini returns:
  • Validate protocol names & chains - currently just LLM searches
  • Match APY within expected tolerance from live feed
    • i.e. more validation of APY CLAIMS

• Max 50% allocation per protocol
• At least 2 different categories
• Max 1 high-risk protocol unless explicitly allowed
• Minimum audit/reputation score

• Periodic refresh every X days
• Auto-rebalance logic
• Notify user via dashboard or assistant when:
  • APYs fall significantly
  • A better opportunity emerges