Bodhi Tree

An open-source AWS-native framework for autonomous DeFi agents integrating MCP for secure, real-time Web3 data and strategy automation

Comment

Inspiration

Many traders and organizations today are looking to leverage AI for 24/7 cryptocurrency trading, especially in the field of DeFi (decentralized finance), which comprises over 10,000 protocols and more than $100B in liquidity. This ecosystem allows users to exchange tokens, lend assets, and increase exposure without borders.

The Model Context Protocol (MCP) serves as the missing link that enables AI agents to access and understand data across different blockchains and protocols. Inspired by the Bodhi Tree — a symbol of enlightenment — we set out to create a framework that helps users gain a complete view of DeFi through intelligent automation.

What it does

Bodhi Tree AI is an autonomous AI-agent framework that bridges AI reasoning with real-world Web3 systems via MCP (Model Context Protocol). It enables users to securely analyze, execute, and optimize DeFi strategies across multiple blockchains and protocols, 24/7.

  • Real-Time Web3 Intelligence

    • Connects AI agents to any MCP server, allowing instant access to multiple blockchains and associated services
    • Aggregates on-chain analytics, liquidity data, and protocol insights

  • Interactive AI Interface

    • Users can chat with AI agents to query on-chain data, monitor portfolios, and receive actionable DeFi insights
    • AI reasoning powered by AWS Bedrock with Claude 4 Sonnet model

  • Autonomous Vault & Strategy Management (Under Development)

    • AI agents can manage DeFi portfolios, execute trades, perform yield farming, and optimize exposure automatically
    • Operates continuously without human intervention

  • Secure Transaction Execution (Under Development)

    • Private keys and sensitive logic are protected using AWS Nitro Enclave
    • Supports both production-grade operations and testing environments on AWS App Runner

How we built it

Bodhi Tree AI was built using a combination of AWS services, MCP servers, and a Next.js dashboard to create a fully autonomous DeFi agent framework.

  • AWS Services

    • Amazon Bedrock for AI reasoning and decision-making using Claude 4 model
    • AWS Lambda + API Gateway for serverless, scalable execution
    • DynamoDB to store agent state, tasks, and chat history
    • CloudWatch for logging, monitoring, and observability
    • AWS CDK for infrastructure as code and automated deployment

  • MCP Integration

    • Connects AI agents to real-time blockchain and DeFi protocol data
    • Supports dynamic addition or removal of MCP servers
    • Provides tool execution and multi-chain analytics

  • Frontend

    • Next.js dashboard for interactive AI chat and DeFi insights
    • Displays real-time analytics, portfolio status, and strategy recommendations

Challenges we ran into

  • MCP Integration: Managing multiple servers and real-time data streams across different blockchains was complex.
  • Autonomous Reasoning: Designing AI agents that act safely and efficiently required careful testing and iteration.
  • Context Retention: Maintaining conversation and task context for multi-turn dialogues and continuous monitoring.

Accomplishments that we're proud of

  • Built an chat interface that can reason and provide insights using MCP servers.
  • Successfully integrated AWS Bedrock for AI reasoning with real-time Web3 data.
  • Designed a serverless architecture that is scalable, secure, and production-ready.
  • Created an interactive Next.js dashboard for AI chat and blockchain monitoring.
  • Enabled real-time access to multiple blockchains and DeFi protocols through MCP servers.

What we learned

  • How to integrate MCP servers with AI models for meaningful DeFi data access.
  • Importance of context management in AI conversations to provide accurate insights.
  • How to design scalable serverless systems using AWS Lambda, API Gateway, and DynamoDB.
  • How AI can augment DeFi analysis even before fully autonomous execution is implemented.

What's next for Bodhi Tree

  • Autonomous vault and strategy execution: AI agents managing portfolios 24/7.
  • Multi-chain support: Ethereum, Kaia, Solana, and other DeFi ecosystems.
  • Advanced risk modeling and portfolio optimization.
  • Expanded AI playground with more tools and data sources.
  • Mobile app and enhanced dashboard for live monitoring and notifications.
  • Open-source ecosystem growth for developers to create custom AI agents.

Built With

Share this project:

Updates