Analyze News, Markets, and Memes with Bias Insights

Helium MCP gives AI coding assistants access to news intelligence, media bias analysis, financial market data, and meme search through 9 tools exposed via the Model Context Protocol. It covers 3.2M+ articles from 5,000+ news sources with 15+ bias dimensions, live stock/ETF/crypto data with AI-generated analysis, and ML-predicted options pricing.

Use it when searching or analyzing news articles with bias-aware context, researching the media bias profile of a specific source or article URL, wanting balanced left/right/center perspectives on a topic, looking up live stock, ETF, or crypto data with AI bull/bear cases, pricing options or evaluating trading strategies, or searching for memes by semantic meaning rather than exact keywords. Two known limitations: tool calls can return empty results for very niche queries, since Helium indexes mainstream and mid-tier sources and hyper-local topics may have limited coverage; and options data is unavailable for tickers without listed options markets, which includes some small-cap stocks and most crypto assets.

The MCP server uses streamable HTTP and needs no authentication:

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

News and media-bias tools: search_news (filter by topic, source, date range, and bias attributes across the article corpus), search_balanced_news (AI-synthesized articles presenting left, right, and center perspectives on a topic), get_source_bias (political lean, factual-reporting score, and 15+ bias dimensions for one source), get_all_source_biases (bias data for all 5,000+ tracked sources in a single call), and get_bias_from_url (full bias analysis on a specific article URL, returning both source-level and article-level indicators). Finance tools: get_ticker (live price, volume, AI-generated bull/bear case, and forecast for a stock, ETF, or crypto ticker), get_option_price (ML-predicted fair value and probability of finishing in-the-money for a specific contract, taking ticker, strike, expiration, and type), and get_top_trading_strategies (top-ranked options strategies for a ticker with risk/reward analysis). The meme tool, search_memes, runs a semantic search by meaning rather than exact keywords.

Works as a Model Context Protocol server compatible with Claude Desktop, Cursor, and Windsurf via a single URL entry in the client's mcpServers config, with no API key or setup beyond that. For building a custom MCP server instead of consuming this one, the docs point to the separate @mcp-builder skill.

AI assistants and their users doing bias-aware news research, cross-referencing source credibility with get_source_bias and search_balanced_news, pairing get_ticker fundamentals with get_option_price or get_top_trading_strategies for actionable trade ideas, or running semantic meme search - all without any authentication setup beyond the MCP config entry.