dolphindb-mcp-server
https://github.com/tradercjz/dolphindb-mcp-serverSkip the database GUI shuffle. This MCP server connects Claude Desktop directly to your DolphinDB cluster, letting you query financial data and run time series analysis through natural conversation.
If you're working with financial data, you know the drill: write query in DolphinDB → copy results → paste into analysis tool → explain findings to stakeholders. This MCP server collapses that entire workflow into Claude conversations.
Ask Claude to "show me the top 10 most volatile stocks from yesterday's trading session" and get immediate results from your DolphinDB cluster. No SQL client switching, no copy-paste gymnastics.
Instant Data Context: Your time series data becomes part of Claude's working memory. Ask follow-up questions, request different time ranges, or pivot analysis directions without losing context.
Natural Language Queries: Instead of remembering DolphinDB syntax, describe what you need in plain English. "Compare trading volumes between tech and finance sectors last week" becomes a conversation, not a SQL exercise.
Seamless Analysis Flow: Pull data, analyze patterns, generate reports, and create visualizations all within a single Claude session. Perfect for exploratory data analysis and ad-hoc research.
Trading Strategy Backtests: "Pull OHLCV data for AAPL from the last 6 months and calculate the Sharpe ratio for a simple moving average crossover strategy."
Risk Analysis: "Show me all positions where daily VaR exceeded 2% in the last quarter, grouped by sector."
Performance Reports: "Generate a summary of our algo trading performance last week, including win rate, average P&L, and maximum drawdown."
Market Research: "Find correlations between commodity prices and related equity sectors over the past year."
The setup is refreshingly straightforward - exactly what you want when connecting critical database infrastructure:
# Install dependencies
uv pip install fastmcp dolphindb dotenv
# Configure your DolphinDB connection
cp .env.example .env
# Edit .env with your DolphinDB host, port, credentials
# Install to Claude Desktop
fastmcp install demo.py
Your DolphinDB data is now accessible through Claude conversations. No additional authentication layers, no complex API management - just direct database access through natural language.
This slots perfectly into existing quant workflows. Your DolphinDB cluster stays exactly where it is, with all your existing security and access controls intact. The MCP server acts as a bridge, not a replacement for your current database setup.
For teams already using DolphinDB for high-frequency trading data or large-scale time series analysis, this eliminates the context switching penalty that kills productivity during data exploration sessions.
The FastMCP framework handles the heavy lifting, so you get reliable Claude Desktop integration without custom MCP protocol implementation.
Ready to query your time series data through conversation instead of SQL clients? Clone it and connect your DolphinDB cluster in under 5 minutes.