OneQAZ
Open-source MCP "context layer" for trading AI: regime detection, self-correcting signals, and macro-to-asset context across 1,100+ symbols.
OneQAZ is an open-source MCP server pitched as "the context layer for financial AI": instead of feeding an agent raw prices, it tells the agent what market regime it is in, which signals are actually working right now, and how macro conditions flow down to individual assets. It covers crypto, US stocks, and Korean stocks across 1,100+ symbols with a 24/7 live cloud API.
Rather than static indicator thresholds, signal weights are self-correcting — adjusted continuously via Thompson Sampling on actual (paper) trade outcomes per regime. It exposes 19 resources and 4 tools for regime status, positions, signals, and trading decisions, plus a macro context chain (global regime → bonds/forex/VIX/commodities → ETF/basket → symbol), Fear & Greed, and an _llm_summary field on every response optimized for agent context windows. Note: its trades are virtual/paper-trading decisions, not real-money order routing, and there is no built-in execution layer.
