The problem Asetta solves

Problem
Asetta addresses 2 major challenges in launching Real-World Asset projects:

• Legal & Compliance Complexity: Bringing real-world assets on-chain requires legal preparation, regulatory compliance, KYC onboarding, and frontend integration. For most teams, this process takes months and often involves hiring multiple costly experts.
• Single-Chain Deployment is Not Enough: Launching on a single chain limits user access and liquidity, while deploying across multiple chains with interoperability protocols is technically complex and difficult to maintain.

RWA projects often require some level of centralization to meet regulatory standards. Unlike IPFS + smart contract solutions, combining AWS Cloud with smart contracts enables RWA creators to issue decentralized tokens while maintaining necessary legal structure and compliance.

Solution

Introducing Asetta, a platform where AI agents specialize and respond in various fields from Legal Agent, AWS Expert Agent to Tokenization Agent and each connected to different MCP servers. By compatible with MCP (Model Context Protocol), RWA creators can setup agents that leverage the extensive MCP server library, now have more 10,000+ MCPs in the market to perform tasks and acquire knowledge of any kind.

The platform comprises several modules as follows:

• Asetta RWA Studio: (https://github.com/tamago-labs/asetta) A desktop application built with Rust + JavaScript, available for download and build from the main repo. Since we aim to support AI agents that natively work with MCP servers, the desktop app is an effective way to run multiple MCP servers. This allows AI agents to access folders containing RWA documents, create legal checklists, and vice versa. Once ready, we can issue RWA tokens on multiple chains, setup Chainlink CCIP configuration with AI agents via Asetta Core MCP tools.
• Asetta Dashboard + Backend: (https://github.com/tamago-labs/asetta-aws-amplify) Built with AWS Amplify, it provides authentication for generating access keys and setting up user and creator profiles. These profiles can be used on both the desktop app and client-side interface that displaying all RWA projects in the system for onboarding investors and trading. It also includes Bedrock AI services for both frontend and desktop apps using Claude models.
• Asetta Smart Contracts: (https://github.com/tamago-labs/asetta-ccip) A set of smart contracts for RWA tokenization and trading deployed on all supported chains and allows AI agents handle token issuance, primary distribution setup, and pricing. We also have secondary market and vault contracts. AI maps everything from your documents to smart contract deployment, reducing errors and improving control.
• Asetta MCP: (https://github.com/tamago-labs/asetta-mcp) An MCP server implementation that enables AI agents to perform tasks using 40+ MCP tools. It connects AI agents to Asetta APIs and smart contracts, facilitating wallet operations and cross-chain configuration and functionality.

How to Use

To create an RWA project, you need to download the desktop app. Currently, we don’t have an executable installer available, so you may need to refer to the Tauri documentation to build it for your OS. Alternatively, you can run it with.:

git clone https://github.com/tamago-labs/asseta
npm install
npm run tauri dev (for development mode)
npm run tauri build (for production build)

Chainlink Integration

We integrate with Chainlink CCIP and have our agent handles the complex process of cross-chain token creation as follows:

• Creates RWA tokens compatible with Chainlink CCIP standards
• Sets up a MintBurnPool for each deployed RWA token and grants roles
• Connect BurnMint pools across chains and validate the entire CCIP setup
• Cross-chain token transfers using Chainlink CCIP from chat

Links:
https://github.com/tamago-labs/asetta-ccip/blob/main/src/RWAToken.sol
https://github.com/tamago-labs/asetta-ccip/blob/main/test/RWAToken.t.sol
https://github.com/tamago-labs/asetta-mcp/blob/main/src/mcp/wallet/configure_ccip_tool.ts

AWS Integration

• The main website and backend use the AWS Amplify serverless stack with Next.js
• Bedrock to provide AI-services to the agent with Claude Sonnet 3.5 (frontend) and Claude Sonnet 4.0 (desktop app)
• Bedrock Knowledge Base syncs with S3 to answer project-specific questions
• Cognito used for secure user authentication across all apps
• Amplify database stores RWA project data and user profiles

Links:
https://github.com/tamago-labs/asetta-aws-amplify/blob/main/amplify/data/resource.ts

Challenges I ran into

Likely the project is a bit complex and dealing with relatively complex use cases. There were many challenges during development, including how to effectively run MCP servers within the application. It turned out that a desktop application was the only viable solution. This allows official AWS MCP servers to run alongside our custom MCP servers, enabling AI agents to process different tasks efficiently. Because of this model, we had to build many different modules that unusual for a typical hackathon project. We need to have a desktop app, AWS webapp and backend, MCP servers, and smart contracts.

While we haven't completed everything we aimed to achieve, we've covered around 70–80% of the core functionality. RWA creators can now chat with AI for project guidance, upload information to the AWS cloud, issue RWA tokens on multiple testnets, and setup vaults for token distributions.

Overall, it was a fun and exciting experience regardless of the outcome. We learned a lot about integrating Web3 and AI on AWS, overcoming real-world challenges, and building solutions that bridge traditional finance with decentralized innovation.