5-agent LLM debate council analyzing macro, fundamental, flow, technical, and skeptic perspectives.
Detects over 10 types of trading biases and emotional patterns in real-time execution.
Real-time VaR calculation, drawdown monitoring, and position sizing optimization.
Ensures all trades meet SEC/FINRA regulations and internal compliance rules.
Processes millions of news articles and social signals to gauge market sentiment.
Recursive learning loop that analyzes past performance to update agent parameters.
Connect any LLM to the TensorTrade ecosystem using our standardized MCP SDK.
{
"mcp_version": "2.0.0",
"agent_config": {
"name": "MarketWatcher",
"capabilities": ["read_price", "analyze_sentiment"],
"permissions": ["execution_read_only"],
"model": "gpt-4-turbo"
},
"context_sources": [
"economic_calendar",
"order_book_l2",
"social_sentiment"
]
}