Hyperliquid Trading Agent

Appearance

Architecture Overview ​

System Components ​

The Hyperliquid Trading Agent consists of several interconnected modules:

Core Modules ​

Signal System ​

  • Orchestrator: Coordinates parallel signal collection
  • Providers: Fetch data from various sources
  • Cache: SQLite-based caching for efficiency
  • Processor: Calculates derived metrics

Decision Engine ​

  • LLM Client: Interfaces with AI models
  • Decision Module: Generates trading decisions
  • Context Builder: Prepares market context for LLM

Execution Layer ​

  • Executor: Places and manages orders
  • Monitor: Tracks positions and performance
  • Market Registry: Manages tradeable assets

Governance System ​

  • Governor: Orchestrates strategy selection
  • Regime Classifier: Detects market conditions
  • Scorekeeper: Tracks strategy performance
  • Tripwires: Implements risk controls

Data Flow ​

  1. Collection: Signal providers fetch data in parallel
  2. Processing: Raw signals are calculated and cached
  3. Analysis: LLM receives processed signals and context
  4. Decision: Trading decision is generated
  5. Execution: Orders are placed on Hyperliquid
  6. Monitoring: Positions are tracked and evaluated
  7. Governance: Performance informs strategy selection

Key Design Principles ​

  • Modularity: Each component has clear responsibilities
  • Async Operations: Parallel signal collection for speed
  • Caching: Minimize API calls and improve performance
  • Extensibility: Easy to add new signals and strategies
  • Observability: Comprehensive logging and monitoring