helium-mcp

Helium MCP Server

Real-time news with bias scoring, live market data, and AI-powered options pricing — all accessible through the Model Context Protocol.

Overview

Helium MCP gives your AI assistant access to:

• 3.2M+ news articles from 5,000+ sources, each scored across 15+ bias dimensions
• Balanced news synthesis — left, right, and center perspectives aggregated into one coherent story
• Live market data — price, IV rank, AI-generated bull/bear cases, and price forecasts
• AI options pricing — proprietary ML models for fair-value estimates and probability ITM
• Meme search — semantic search across viral memes with OCR text and captions

50 free queries included. No sign-up required.

Quick Start

Cursor

Add to your Cursor MCP settings:

{
  "mcpServers": {
    "helium": {
      "url": "https://heliumtrades.com/mcp"
    }
  }
}

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "helium": {
      "type": "streamable-http",
      "url": "https://heliumtrades.com/mcp"
    }
  }
}

Windsurf

Add to your Windsurf MCP settings:

{
  "mcpServers": {
    "helium": {
      "serverUrl": "https://heliumtrades.com/mcp"
    }
  }
}

GitHub Copilot

Add to your .github/copilot-instructions.md or MCP config:

{
  "mcpServers": {
    "helium": {
      "type": "streamable-http",
      "url": "https://heliumtrades.com/mcp"
    }
  }
}

Any MCP Client

Connect to the streamable HTTP endpoint:

https://heliumtrades.com/mcp

Example Output

"What's the bull and bear case for Bitcoin?"

get_ticker("BTC")
→ Bitcoin: $71,040.01
  Bull case: "Institutional spot demand resumes... squeeze toward mid/upper-80ks to low-100ks"
  Bear case: "Higher-for-longer rates, persistent ETF outflows... slide into 40–55k plausible"
  Forecast: +1.8% over 37 days (range: -1.2% to +5.0%)
  IV Rank: available
  Options strategies: long vol + short vol packs included

"How biased is CNN?"

get_source_bias("CNN")
→ Emotionality: 7/10
  30+ bias dimensions scored (political lean, fearfulness, opinion, oversimplification...)
  Signature phrases: "trump threatens", "mortgage rates", "nuclear weapons"
  Similar sources: NBC, PBS, AP, Newsweek, CBS
  Per-article bias breakdown on recent stories included

"What are the best options trades right now?"

get_top_trading_strategies()
→ AI-ranked short vol + long vol setups
  Each includes: ticker, price, bull/bear case, 5 probability-weighted outcomes,
  full option pack (strike, expiry, Greeks, ML fair value, probability ITM)

"Give me NVIDIA's full analysis"

get_ticker("NVDA")
→ NVIDIA: $183.91
  5 probability-weighted scenarios (38% mean-revert, 25% +10-25%, 20% -5-12%...)
  Forecast: +4.2% over 37 days (range: -2.6% to +8.9%)
  IV rank, volatility surface, term structure, return profile charts
  27 time-horizon uncertainty visualizations (1-day to 995-day)

"Find me stock market crash memes"

search_memes("stock market crash")
→ Babylon Bee: "Kamala Desperately Googles 'What Is A Stock Market Crash'" (32K likes)
→ "Everyone: 'omg the stock market is crashing!' / 500 cigarettes adapter" (437K likes)
  Full OCR text, source, date, image URLs included

Tools

search_news

Full-text search across 3.2M+ articles from 5,000+ sources. Filter by source, category, date, and minimum shares. Returns bias scores, AI summaries, and context.

Parameters:

Parameter	Type	Required	Default	Description
query	string	Yes	—	Search keywords
limit	integer	No	20	Max results (1–100)
source	string	No	—	Filter by source name (e.g. "CNN", "Reuters")
category	string	No	—	One of: trending, tech, markets, politics, business, science, memes
days_back	integer	No	720	Only include articles from the last N days
min_shares	integer	No	-1	Minimum total social shares
sort	string	No	"rank"	One of: rank (relevance), date (newest), shares (most shared)

search_balanced_news

AI-synthesized articles that aggregate multiple sources into balanced coverage. Each story includes a summary, takeaway, evidence breakdown, potential outcomes, and relevant tickers.

Parameters: