MarketDX

The financial impact graph.
A structured, explained map of what financial news is about and who it touches — the why (causal channel), the story's lean, and the ripple — across stocks, commodities, FX, crypto, and private companies. A research & context layer, not a trading signal.

▶ Live playground · Docs & pricing · Sample dataset ↓

---

Most feeds give you a headline and a sentiment score. MarketDX labels every affected entity on every event, across many dimensions — 10 causal channels (demand · supply · regulation · tariff · competition · pricing · capital · technology · geopolitics · monetary), per-entity direction & relevance, epicenter and ripple, across five asset classes including non-listed companies ticker feeds structurally can't see. It's a layer for understanding and structuring the news flow — a research, screening and feature input — not a predictive price signal (news is largely priced in by the time it's public).

What a signal looks like

One headline → every entity it touches, each labeled with the lean (+/−) and the why. Take "AI-driven memory shortage forces Apple, Samsung, Microsoft… to raise device prices":

entity		lean	why (aspect)
Apple	stock	−	pricing — chip costs squeeze margins
Microsoft	stock	−	pricing
Nintendo	stock	−	pricing
Samsung	stock	+	demand — it makes the memory

Same event, opposite outcomes — because the why is labeled, not just a sentiment score. And it spans five asset classes including private companies (OpenAI, Anthropic, …) that ticker feeds structurally can't see, plus the ripple into non-obvious names. _{↑ a real row from the free sample dataset — no signup needed.}

What's here

📊 datasets/	A free sample of the impact graph — 4,000+ labeled signals, every angle the playground answers. Start here.	✅ available
🐍 python/	The official Python SDK — the graph in three lines. pip install marketdx	✅ available
🔌 mcp/	MCP server — give your AI agent the financial impact graph.	🔜 soon

New here? → browse the sample dataset (no signup), then try it live.

Quickstart

See the data — no signup: browse the sample dataset or the live playground.

Use it in code — the official Python SDK:

pip install marketdx

from marketdx import MarketDX

mdx = MarketDX(api_key="…")                 # free key at marketdx.lab.ai
for s in mdx.news(megatrend="ai-power", impact="indirect"):
    print(s.title, [(e.name, e.impact.net_direction) for e in s.entities])

Full SDK docs → python/.

Free to explore — the sample dataset and playground need no signup, and the API has a free tier. Paid plans unlock full & live coverage — see pricing.

License

Sample data under CC BY 4.0; code (SDK, MCP) under MIT. See each directory. Built by MarketDX — democratizing financial data.