Architecture

Key components

• Intelligence Engine -

  • LLM - Processes and interprets user queries using advanced natural language capabilities to generate structured actionable intents

  • Model Context Protocol - Bridges the LLM’s output and on-chain execution logic using micro-services that integrate rule-based heuristics with AI-driven decision trees to ensure precise, context-aware blockchain operations

  • Path Finder - Maps out the most effective sequence of actions for executing blockchain intents based on the processed query

• Settlement Engine -

  • Home Chain - Contains set of smart contracts that power dBOOK

  • Operators - Execute tasks on behalf of users while funds remain in user accounts until fulfilment. They include solvers, fillers (for cross-chain trades), KYC providers (for compliance), and CEXs (for off-exchange settlements)

  • Validators - Secure the network by controlling MPC endpoints on different chains and ensuring operators act honestly. dBOOK is consensusless, using optimistic validity proofs to ensure security without sacrificing speed

  • Accounts - Any address (user or smart contract) can hold funds on any chain and define on-chain policies for compliance such as control what assets can be sent, reject incoming transactions or even delegate such screening to third parties

• Recommendation Engine -

  • Protocol & Metadata Registry - An up-to-date database of blockchain protocols and their metadata, enabling accurate matching with user goals

  • Context Module - Analyzes transient data from the Intelligence Engine to dynamically align user actions with the most suitable protocols