Architechure
Last updated
Last updated
Operari represents an innovative approach within the decentralized finance landscape, functioning as an advanced Universal Orchestration Platform that bridges traditional web interfaces with blockchain ecosystems. It's architected as an intelligent, comprehensive system that seamlessly integrates AI-Powered Agents with both Web Navigation Tools and DeFi Protocol Interactions to create a unified experience across digital boundaries.
Our core mission is to serve as the Universal Cognitive Interface for DeFi operations, transforming complex multi-platform workflows into intuitive, natural language experiences. Where traditional interfaces require technical knowledge and multiple steps, Operari leverages advanced AI to make digital interactions profoundly more accessible, automating complexity and optimizing execution across platforms.
At a high level, Operari's architecture can be understood through interconnected layers:
This foundation layer powers Operari's contextual intelligence through:
Internet Sources: Real-time data collection from websites, repositories, and other online resources ensures the system remains current with the latest information.
Knowledge Base: Structured information stored in both Vector and Graph databases provides semantic understanding and relational context, enabling intelligent decision-making.
This comprehensive knowledge architecture allows Operari to understand complex concepts, interpret user intent, and formulate appropriate execution strategies.
Sitting at the heart of the ecosystem, the Operari Orchestrator Agent acts as the central cognitive hub. It receives natural language inputs from users, determines intent, and coordinates execution across specialized agents.
It manages session state, orchestrates multi-step workflows, and maintains contextual awareness throughout interaction chains, ensuring coherent experiences across different platforms and protocols.
This layer translates abstract user desires into concrete action plans that can be executed by specialized agents.
This layer comprises specialized AI agents with domain-specific capabilities:
CUA (Computer User Agents): These agents interact with traditional web interfaces, performing actions like scrolling, clicking, inputting text, and capturing screenshots. They effectively serve as an AI-powered bridge to existing web applications, eliminating the need for specialized integrations.
DeFAI Agents (Decentralized Finance AI): These specialized agents handle all blockchain interactions, from initiating swaps via KittenSwap to managing complex transaction workflows and connecting with frontend interfaces for user signatures.
Both agent types leverage RAG (Retrieval-Augmented Generation) capabilities to access the Knowledge Layer when needed, ensuring decisions are informed by current, relevant information.
The Auth State Storage system securely manages authentication across platforms:
Stores and manages authentication tokens for platforms like X (formerly Twitter) and various applications
Enables seamless cross-platform operations without requiring repeated user authentication
Provides a secure foundation for delegated actions while maintaining appropriate security boundaries
This layer handles the actual execution of transactions and operations:
User Wallet: Securely manages user identities and private keys for blockchain transactions
Frontend/UI: Provides visual feedback and interaction points for users, including dashboards and transaction signing interfaces
Protocol Integration: Connects directly with various blockchain protocols (Swap, Lending, DEX) to execute financial operations
The execution layer ensures that the plans formulated by the Orchestration Layer and implemented by the Capability Layer are properly completed and verified.
Operari's innovative architecture creates a seamless bridge between human intent and digital execution, eliminating technical barriers and enabling intuitive interactions with both traditional web interfaces and decentralized protocols. The system continuously learns from interactions, improving its capability to understand user intent and optimize execution strategies over time.