Showcase: Chart Library MCP server for financial pattern intelligence

[grahammccain]

Showcase

Chart Library MCP is a production MCP server for historical chart pattern intelligence, built with Python. Sharing it here as a real-world example of a domain-specific MCP server that might be useful for the fastmcp community.

What it does

Chart Library provides 24M+ pre-computed chart pattern embeddings across 19K stock symbols and 10 years of market data. The MCP server exposes 19 tools:

Core Search (7 tools)

• search_pattern — Find the 10 most similar historical chart patterns by ticker + date
• search_by_image — Upload a chart screenshot to find similar patterns
• get_forward_returns — 1/3/5/10-day forward return distributions
• get_pattern_details — Technical pattern detection with confidence scores
• And more...

Market Intelligence (7 tools)

• get_sector_rotation — Sector ETF rankings by relative strength
• get_anomaly_detection — Flag unusual pattern deviations
• get_correlation_shift — Stocks decorrelating from SPY
• get_volume_profile — Intraday volume vs historical average
• And more...

Trading Intelligence (4 tools)

• get_regime_win_rates — Win rates filtered by current market regime
• get_pattern_degradation — Are signals losing accuracy recently?
• get_exit_signal — Pattern-based exit recommendations
• get_risk_adjusted_picks — Sharpe-like scoring per pick

Install & use

pip install chartlibrary-mcp

{
  "mcpServers": {
    "chartlibrary": {
      "command": "python",
      "args": ["-m", "chartlibrary_mcp"]
    }
  }
}

Stats

• Glama: A A A score
• PyPI: chartlibrary-mcp v1.1.0
• Free tier: 200 API calls/day, no credit card

Would love feedback from the fastmcp team on the server design. Happy to contribute it as an example if you're building out a gallery of MCP servers.