market-data-mcp
Provides real-time market data tools (quotes, news, earnings calendar, watchlist scanner, and composite analysis) for AI agents via Finnhub, with optional Alpaca broker integration and graceful degradation.
README
market-data-mcp
Real-time market data: quotes, news, earnings calendar, and a watchlist scanner for AI agents — backed by Finnhub.
Gives any MCP client (Claude Desktop, Claude Code, agents) clean, uniform market data tools that fail gracefully. A missing or premium-gated source returns "no data", never a fabricated value. Markdown output by default, JSON on demand.
Tools
| Tool | What it does | Source | Free tier |
|---|---|---|---|
market_quote |
Real-time price, % change, high/low/open/prev-close | Finnhub /quote |
Yes |
market_news |
Top 5 headlines — company-specific or general market | Finnhub /company-news, /news |
Yes |
market_calendar |
Earnings calendar for the next N days (with EPS/revenue estimates) | Finnhub /calendar/earnings |
Yes* |
market_scan |
Watchlist scanner: rank up to 25 tickers by absolute % change | Finnhub /quote x N |
Yes |
market_analyze |
Composite: momentum score + news catalyst + earnings proximity | All of the above | Yes |
broker_positions |
Read open positions from Alpaca paper account | Alpaca Paper API | Optional** |
*Calendar tested live on 2026-06-14: accessible on Finnhub free tier.
**broker_positions requires ALPACA_API_KEY + ALPACA_SECRET_KEY in .env. Degrades gracefully without them.
Every tool returns Markdown (human-readable, default) or JSON (response_format="json") for programmatic use.
Live demo output (2026-06-14)
market_quote AAPL:
AAPL: $291.13 (-1.52% -$4.50) prev close $295.63
market_scan AAPL,MSFT,NVDA,TSLA:
Scanned 4/4 symbols. Top mover: TSLA +1.82%
market_calendar days=7:
20 earnings events in the next 7 days: ACN, KR, MEI ...
market_analyze AAPL:
Signal: Lean negative [......##|........] -30/100 confidence 60%
AAPL: $291.13 -1.52% → Lean negative [TRIM]
Catalyst: Apple's iOS 27 surprise could change the AI narrative
Quick start
git clone <your-repo-url> market-data-mcp
cd market-data-mcp
python -m venv .venv
.venv\Scripts\activate # Windows
pip install -r requirements.txt
copy .env.example .env # edit: add FINNHUB_API_KEY
python tests/test_smoke.py # live test all 6 tools
Get a free Finnhub key at: https://finnhub.io/register
Claude Desktop config
Add this to claude_desktop_config.json
(%APPDATA%\Claude\ on Windows, ~/Library/Application Support/Claude/ on macOS),
then restart Claude Desktop:
{
"mcpServers": {
"market-data-mcp": {
"command": "python",
"args": ["-m", "market_data_mcp"],
"cwd": "C:/path/to/market-data-mcp"
}
}
}
Use the system Python path if you are not using a venv:
{
"mcpServers": {
"market-data-mcp": {
"command": "C:/Users/YourName/AppData/Local/Python/pythoncore-3.14-64/python.exe",
"args": ["-m", "market_data_mcp"],
"cwd": "C:/path/to/market-data-mcp",
"env": { "PYTHONUTF8": "1" }
}
}
}
Optional: Alpaca broker source
The broker_positions tool reads your Alpaca paper account. To enable it:
- Go to https://app.alpaca.markets/ → Paper Trading → API Keys
- Add to
.env:ALPACA_API_KEY=your-key-here ALPACA_SECRET_KEY=your-secret-here - Restart the server.
Without these keys, broker_positions returns a graceful "keys not set" message.
Why it's built this way
- Uniform result contract. Every source returns the same
Resultshape (source, ok, summary, data, score, confidence, error). An LLM can reason across all tools without parsing N formats. - Graceful degradation. A source with no data, a premium-gated endpoint, or a missing key returns
Result.failed(...)— never a fabricated value. Zero-price from Finnhub for unknown symbols is treated as "no data", not a quote. - Scored analysis. The composite
market_analyzetool emits a -100..+100 directional score with confidence, so an agent can triage without reading prose. - TTL cache. Per-source in-process cache avoids hammering rate-limited APIs when an agent calls several tools in one turn (e.g. scan + analyze in sequence).
Disclaimer
For research and educational use only. Data comes from Finnhub and Alpaca and may be delayed or incomplete. Never use automated market data for financial decisions without independent verification.
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.