oneqaz-trading-mcp

The context layer for financial AI.

Your AI agent shouldn't just see prices — it should understand what regime the market is in, which signals are actually working right now, and how macro flows down to individual assets.

OneQAZ provides this as a single MCP endpoint. Crypto, US stocks, Korean stocks. 1,100+ symbols. 24/7 live.

Keywords: MCP, trading, signals, market analysis, regime, portfolio, sentiment, technical analysis, crypto, stocks, Fear & Greed, cross-market, Claude, model context protocol

Why OneQAZ

Financial data APIs are everywhere. Market intelligence is not.

Signal weights are adjusted continuously based on actual trade outcomes per regime via Thompson Sampling — not static indicator thresholds. Every response includes an _llm_summary field optimized for AI consumption.

What your AI gets

• Regime detection: Is the market trending, ranging, or volatile? Per-market and global
• Self-correcting signals: 1,100+ symbols scored by Thompson Sampling on actual trade outcomes
• Macro context chain: Global regime → bonds/forex/VIX/commodities → ETF/basket → individual symbol
• External context: News events, fundamentals, cross-market correlation — pre-processed for LLM consumption
• 19 Resources + 4 Tools: Stateless HTTP, compatible with any MCP client
• _llm_summary on every response: Human-readable text summary optimized for AI agent context windows

Market Coverage

All symbols are monitored 24/7 with automated signal generation, regime detection, and virtual trading.

Quick Start

Option 1: Live API — no install needed

Real-time data, updated every minute.

{
  "mcpServers": {
    "oneqaz-trading": {
      "url": "https://api.oneqaz.com/mcp"
    }
  }
}

Ask Claude: "What's the current market regime?"

Option 2: Local (demo data)

pip install oneqaz-trading-mcp
oneqaz-trading-mcp init    # creates sample SQLite databases
oneqaz-trading-mcp serve   # starts at http://localhost:8010

• Swagger UI: http://localhost:8010/docs
• MCP endpoint: http://localhost:8010/mcp

Then connect from Claude:

{
  "mcpServers": {
    "oneqaz-trading": {
      "url": "http://localhost:8010/mcp"
    }
  }
}

Use Cases

1. Give your AI agent market awareness

Connect OneQAZ and your agent understands market context without you building the pipeline:

# Your agent reads regime + signals + macro in one call
context = mcp.read("market://crypto/unified")

# Or go granular
regime = mcp.read("market://crypto/status")          # what phase is the market in?
signals = mcp.call("get_signals", market_id="crypto", min_score=0.7)  # what's working now?
macro = mcp.read("market://global/summary")           # what's driving this from above?

# Feed to your agent's decision layer
prompt = f"""
  Regime: {regime}
  High-confidence signals: {signals}
  Macro context: {macro}

  Recommend portfolio action.
"""

2. Build a regime-aware trading system

Your AI reacts differently based on market state — no hardcoded rules:

regime = mcp.read("market://us_stock/status")
structure = mcp.read("market://us_stock/structure")

if regime["regime"]["stage"] == "volatile":
    signals = mcp.call("get_signals", market_id="us_stock", action_filter="DEFENSIVE")
else:
    signals =

... [View full README on GitHub](https://github.com/oneqaz-trading/oneqaz-trading-mcp#readme)