> For the complete documentation index, see [llms.txt](https://rapidchain.gitbook.io/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://rapidchain.gitbook.io/docs/white-paper/technical-architecture.md). # Technical Architecture #### 2.1. The Three-Layer Stack Rapid Chain is a **high-assurance EVM network**. Instead of treating speed and certainty as a trade-off, it separates concerns so that high-frequency execution and final settlement happen in the **same block**. **Layer 1: User Interface** * Rapid Wallet for assets, tokens and dApp interactions * RapidSwap (AMM) and Rapid Order (CLOB) trading interfaces * Token Pilot and RocketPad for no-code token creation and launches * No-Code Agent Builder for AI agent deployment * Rapid Scan explorer for real-time monitoring and analytics **Layer 2: Execution Layer** * **Full EVM** for standard smart contracts * **RAda VM** for safety-critical, formally verified operations * **AI Agent Runtime** for autonomous decision-making and execution * **Micro-Payment Hub** for high-frequency agent-to-agent transactions * **RAPALGO Engine** for cross-exchange intelligence and derivatives signals **Layer 3: Finality Layer** * Deterministic BFT consensus: validators co-sign every block before it is appended * Single-block finality, with no forks to resolve and no reorganisation window * Validator set governed by an on-chain smart contract * Protocol-level permissioning, available on demand **Sovereignty & Control:** Every participant keeps full control of their own keys, nodes and AI models. No state change can be forced without a valid, quorum-signed transaction, and no finalised block can ever be rolled back. ![Speed Meets Certainty: execution and finality in the very same block.](https://4061171286-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FCAkaYak0f6hx6Oc1NfOP%2Fuploads%2Fgit-blob-fd199ccc317b33f0aad2d674e82f677e8261f25e%2Fspeed-meets-certainty.png?alt=media) #### 2.2. AI Agent Runtime The **AI Agent Runtime** is Rapid Chain's specialized environment for autonomous economic agents. Unlike traditional smart contracts that execute predefined logic, AI agents operate as **persistent, stateful entities** that can: * **Perceive** market conditions via oracle feeds and API integrations * **Decide** optimal actions using machine learning models or rule-based systems * **Execute** transactions autonomously within defined risk parameters * **Settle** payments instantly via the Micro-Payment Hub **Agent Architecture:**
Agent Identity, Agent Wallet, Execution Sandbox and Settlement Connector
| Component | Responsibilities | | ------------------------ | ----------------------------------------------------------------------- | | **Agent Identity (DID)** | Reputation scoring, KYC/AML verification, spending limits & rules | | **Agent Wallet** | Auto-balance management, micro-payment channels, subscription billing | | **Execution Sandbox** | Python / Node.js runtime, WASM for compiled strategies, metered compute | | **Settlement Connector** | Single-block on-chain settlement, compliance enforcement, audit trail | **Use Cases:** * **Risk Agents:** Real-time collateral monitoring and margin call automation * **Trading Agents:** Order placement on Rapid Order, liquidity management on RapidSwap and hedging signal execution * **Payment Agents:** Automated invoice settlement and recurring payments in SETUSD * **Oracle Agents:** External data validation and on-chain attestation **Technical Implementation: Agent Execution Flow** The following structure defines how an AI agent processes decisions and settles them on Rapid Chain: ```python # Rapid Chain AI Agent Runtime - Python SDK (illustrative) from rapidchain import Agent, Wallet, Settlement class RiskMonitoringAgent(Agent): """ Autonomous agent for real-time collateral monitoring and automated margin call execution. """ def __init__(self, config): super().__init__( name=config["agent_name"], wallet=Wallet.create( custody="mpc", spending_limit=config["max_daily_spend"] ), rules={ "kyc_required": True, "audit_level": "full", "compliance_hooks": ["aml_check", "sanctions_screen"] } ) # Settle directly on Rapid Chain; final in a single block self.settlement = Settlement( network="rapid-chain", settlement_asset="SETUSD" ) def on_price_threshold(self, market_signal): """ Triggered when collateral value drops below threshold. Executes an automated margin call settled on-chain. """ # 1. PERCEIVE: Validate market data if not self.validate_oracle(market_signal): self.log("Invalid oracle data - aborting") return False # 2. DECIDE: Calculate required collateral adjustment collateral_deficit = self.calculate_deficit(market_signal) if collateral_deficit > 0: # 3. EXECUTE: Prepare margin call instruction margin_call = { "collateral_vault": self.config["collateral_vault"], "asset": "SETUSD", "value": collateral_deficit, "instruction_type": "MARGIN_CALL", "compliance_proof": self.generate_compliance_proof() } # 4. SETTLE: The receipt is returned once the block is final receipt = self.settlement.execute(margin_call) self.log(f"Margin call settled in block {receipt.block_number}: {receipt.tx_hash}") return receipt.tx_hash return None def validate_oracle(self, signal): """Multi-source oracle validation for price feeds.""" consensus = self.query_oracle_consensus( sources=["chainlink", "internal", "rapid_oracle_agent"], threshold=0.95 # 95% agreement required ) return consensus.valid # Agent deployment and activation if __name__ == "__main__": # Initialize agent with institutional parameters agent_config = { "agent_name": "RepoRiskMonitor_01", "max_daily_spend": 100000, # USD "collateral_vault": "0x..." # Collateral contract on Rapid Chain } agent = RiskMonitoringAgent(agent_config) # Register agent in Rapid Chain Marketplace agent.deploy( marketplace="rapid_chain_main", fee_share=0.70, # 70% to agent owner subscription_price=500 # Monthly USD ) # Start autonomous monitoring agent.start_listening( trigger="collateral_ratio", threshold=1.05 # 105% collateral threshold ) ``` **Key Runtime Features:** | Feature | Implementation | | --------------------------- | -------------------------------------------------------------------------------------------------------------------- | | **Sandbox Isolation** | Each agent runs in a containerized environment with resource limits | | **State Persistence** | Agent memory and learning state stored in an encrypted local database | | **Single-Block Settlement** | All financial actions settle via `settlement.execute()`; a receipt means the block is final, with no rollback window | | **Compliance Enforcement** | KYC/AML hooks execute before any state-changing operation | | **Metered Execution** | Compute and storage costs tracked in real time, billed in $RAPID | #### 2.3. Deterministic Finality and Contract-Governed Validators The core of Rapid Chain's reliability is its **finality engine**. **Single-Block Finality:** Rapid Chain runs a deterministic BFT consensus protocol. Validators co-sign each block before it is appended; once the quorum has signed, the block is final. There are no forks to resolve and no "wait N confirmations". A deposit, a trade or a margin call is final the instant its block exists. **Contract-Governed Validator Set:** The list of validators lives in an on-chain smart contract. Adding or removing a validator is itself a transaction: public, timestamped and permanently auditable on Rapid Scan. **Permissioning on Demand:** The protocol can enforce node- and account-level allowlists natively, without a custom fork or sidecar. Regulated participants can operate in controlled zones while the network as a whole stays open. #### 2.4. Hybrid Execution Environment (EVM & RAda) ![RAda Layer](https://rapidchain.io/images/6.png) Rapid Chain provides a **dual-engine virtual machine** to serve both Web3 developers and high-assurance institutional systems. **Full EVM Compatibility:** Standard Solidity smart contracts deploy unchanged, enabling rapid migration of existing DeFi protocols while benefiting from single-block finality. RapidSwap, Rapid Order and SETUSD all run on the EVM layer today. **RAda High-Assurance VM:** For safety-critical operations, such as collateral netting, repo margin calls and high-value settlements, RAda provides aerospace-grade formal verifiability. Every instruction has predictable, deterministic state transitions with mathematical proof of correctness. **Selective Usage:** Developers choose the appropriate environment: EVM for rapid innovation, RAda for institutional-grade safety. #### 2.5. Interoperability: SETBridge Rapid Chain connects outward without compromising its own finality. **SETBridge** links BNB Chain liquidity to Rapid Chain: * Every USDT locked on BNB Chain mints **SETUSD 1:1** on Rapid Chain * Each mint lands in a single, final block and is verifiable on Rapid Scan * Once bridged, SETUSD trades natively on Rapid Order and in RapidSwap pools **Key Characteristics:** * **Transparent:** Every mint is publicly visible on Rapid Scan * **Final:** No ambiguous "in-between" states on the Rapid Chain side * **Secured contract:** SETUSD is built to OpenZeppelin v5 standards with role-based access control and an emergency pause mechanism #### 2.6. Network Parameters | Feature | Specification | | -------------------------- | ----------------------------------------------------------------------------------- | | **Consensus Mechanism** | Deterministic BFT with a contract-governed validator set | | **Finality** | Single block, with no confirmation wait and no reorganisation | | **Privacy Model** | Public ledger by default; protocol-level permissioning on demand; staged ZK roadmap | | **State Storage** | Lean, flat, pruned state layout for fast sync on modest hardware | | **Interoperability** | SETBridge (BNB Chain → Rapid Chain) | | **Smart Contract Support** | EVM (Solidity) + RAda (High-Assurance) | | **Developer APIs** | JSON-RPC, WebSocket subscriptions, GraphQL | | **Observability** | Prometheus / OpenTelemetry metrics on every node | | **AI Agent Runtime** | Python / Node.js / WASM with isolated sandbox | | **Micro-Payments** | Batched via the Micro-Payment Hub |