moon-dev-trading-agents

Multi-exchange AI crypto trading agent orchestration is handled by a main Python orchestrator script that coordinates 48+ specialized agents across Hyperliquid, BirdEye, Asterdex, and Extended Exchange. A modular codebase routes signals, risk checks, and execution pipelines autonomously across the connected platforms.

Yes, you can use different LLM providers for crypto trading agents through an LLM abstraction layer. A ModelFactory module unifies access to Claude, GPT-4, Gemini, and other providers, allowing agents to switch models without altering the core orchestration logic.

Backtesting autonomous trading strategies across multiple crypto exchanges is executed within the agent orchestration framework. The system coordinates backtesting workflows alongside risk checks and execution pipelines, evaluating strategy performance using historical market data before live deployment.

Python 3.10 or higher is required to run an autonomous multi-agent crypto trading system. The advanced implementation relies on a modular codebase structure with dedicated directories for agents, models, strategies, and data to function properly.

Risk limits and exchange connections for AI trading agents are configured in the src/config.py file. This central configuration module allows teams to specify target exchanges, select the active AI provider, and set risk thresholds for the orchestrator.

Running individual AI trading agents separately, such as trading_agent.py or risk_agent.py, allows developers to debug, test, and extend specific agent behaviors in isolation. This modular approach isolates execution issues before integrating them into the full multi-exchange orchestrator.