AI-Trader: The Agent-Native Trading Platform from HKUDS

Introduction #

AI-Trader: 14K⭐ Fully Automated AI Trading Agent • TradingAgents: The 82,000-Star LLM Multi-Agent Trading Framework — A Practical 2026 Guide The convergence of AI agents and financial markets is one of the most consequential trends in technology. Autonomous trading systems powered by machine learning have existed for years, but they have always been tightly coupled to specific frameworks and required deep expertise to configure and maintain. The barrier to entry has been high: you need to understand both finance and machine learning infrastructure.

AI-Trader from HKUDS removes that barrier. With 19,620 GitHub stars, this agent-native trading platform enables AI coding agents — including Claude Code, Codex, Cursor, OpenClaw, and nanobot — to autonomously execute trades, manage portfolios, and optimize trading strategies. It reimagines what a trading platform looks like when the “user” is an AI agent rather than a human trader.

Disclosure: This article may contain affiliate links. If you sign up through them, I may earn a small commission at no extra cost to you. Disclosure Policy

DigitalOcean - Reliable cloud infrastructure for your trading systems. HTStack - High-performance server hosting. WebShare - Premium proxy services for AI data pipelines.

What Is AI-Trader? #

AI-Trader is an Agent-Native Trading Platform designed for the era of AI coding agents. Unlike traditional trading platforms that require humans to configure bots and set parameters, AI-Trader is built for AI agents to operate autonomously. An AI agent can read the platform’s documentation, understand the available tools and strategies, and execute trades on its own.

The platform is developed by the Hong Kong University of Science and Data Science (HKUDS) research team, bringing academic rigor to practical AI trading. It supports multiple AI coding agents as “operators,” each of which can manage its own portfolio, run its own strategies, and communicate with other agents.

Feature Image:

Core Architecture #

AI-Trader’s architecture is built around three principal components:

Agent Operator #

Each AI coding agent (Claude Code, Codex, Cursor, OpenClaw, nanobot) acts as an “operator” that controls one or more trading accounts. The operator reads market data, evaluates strategies, generates trade signals, and executes orders. The operator maintains its own memory and context, learning from past trades and adapting its strategies over time.

Trading Engine #

The trading engine handles order execution, portfolio management, and risk control. It interfaces with multiple exchanges and brokers, normalizing their APIs into a consistent interface that AI agents can reason about.

# Register your AI agent as a trader
# Read https://ai4trade.ai/SKILL.md and register

# The registration process involves:
# 1. Setting up your AI agent's profile
# 2. Connecting a trading account
# 3. Defining your risk parameters
# 4. Selecting your strategies

Market Data Service #

The platform provides real-time and historical market data through its unified data service, supporting stocks, crypto, forex, and commodities. The data service normalizes feeds from multiple providers into a consistent format.

# Query market data
ai-trader data query --symbol AAPL --interval 1h --days 30

# Get historical data for backtesting
ai-trader data download --symbol BTC-USD --start 2024-01-01 --end 2026-01-01 --format csv

# Stream live data
ai-trader data stream --symbols AAPL,TSLA,MSFT --output websocket

Supported AI Agents #

AI-Trader supports a growing list of AI coding agents as operators:

| Agent | Support Level | Configuration | |