Financial News Screening with MCP Search

AI-powered financial news screening via MCP search tools costs nothing on free tiers and replaces expensive news subscriptions for personal use. Set up Claude Code or a local LLM with a search MCP server, query topics in natural language, and get structured results from financial news sources, Reddit discussions, and market data pages.

MCP search for news screening

MCP (Model Context Protocol) lets AI tools call search APIs as native tools. Instead of opening a browser and searching manually, ask your AI assistant to search for financial news topics. The AI reads the results, filters for relevance, and summarizes key points.

Setup options

• Tavily MCP: 1,000 free searches/month. Returns summarized content that bypasses paywalls in some cases. Community-maintained MCP server.
• Scavio MCP: 250 free credits/month. Hosted at mcp.scavio.dev/mcp. Returns structured JSON from Google, Reddit, YouTube. Multi-platform coverage in one endpoint.
• Brave Search MCP: ~1,000 free queries/month ($5 free credits). Raw search results. Several community MCP implementations.

Configure Claude Code for news screening

{
  "mcpServers": {
    "scavio": {
      "type": "url",
      "url": "https://mcp.scavio.dev/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Daily screening workflow

import os, requests
from datetime import date

H = {"x-api-key": os.environ["SCAVIO_API_KEY"],
     "Content-Type": "application/json"}
BASE = "https://api.scavio.dev/api/v1/search"

WATCHLIST = [
    "ASX market news today",
    "RBA interest rate decision 2026",
    "lithium mining stocks australia",
    "AI sector earnings report",
]

def daily_screen():
    today = date.today().isoformat()
    print(f"Financial News Screen - {today}")
    for topic in WATCHLIST:
        resp = requests.post(BASE, headers=H,
            json={"query": topic, "country_code": "au"})
        results = resp.json().get("organic_results", [])[:3]
        print(f"\n{topic}:")
        for r in results:
            print(f"  {r.get('title', '')[:70]}")
            print(f"  {r.get('link', '')}")
    cost = len(WATCHLIST) * 0.005
    print(f"\nCost: {cost:.3f} ({len(WATCHLIST)} queries)")

daily_screen()

Reddit as a leading signal

Financial subreddits often discuss market events before mainstream news picks them up. Adding Reddit as a search platform catches discussion threads about earnings surprises, regulatory changes, and sector rotation before they appear in traditional news feeds.

Cost comparison

Bloomberg Terminal: $24,000/year. Financial Times: $500/year. The MCP search approach with 10 daily queries: $18.25/year at $0.005/query (or free on Tavily's 1,000/month tier). Obviously not the same depth, but for personal screening and staying informed, the signal-to-cost ratio is hard to beat.