Mcp Portfolio Ideas

What is MCP Portfolio Ideas?

MCP Portfolio Ideas is a hosted remote MCP server that enables LLMs to create, analyze, and optimize equity portfolios using structured tools. It integrates with portfolio optimization libraries (PyPortfolioOpt, RiskfolioLib) and financial data via yfinance, and is intended for users who want reproducible, strategy-aware portfolio analysis inside chat.

How to use MCP Portfolio Ideas?

Configure your MCP client with the SSE endpoint https://mcp-portfolio-ideas.metapartner.ai/mcp/sse/ and set the header X-API-Key to the guest key provided in the README. Use the tool catalog to create portfolios, inspect metrics, compare against benchmarks, and run optimizations.

Key features of MCP Portfolio Ideas

• Create equity portfolios from symbols or weight maps
• Inspect risk, concentration, sector exposure, tail risk
• Compare portfolios against benchmark portfolios
• Optimize portfolios using a suite of predefined strategies
• Guest key provides full access to all tools and strategies
• Shared workspace cleans up nightly (guest mode)

Use cases of MCP Portfolio Ideas

• Build a portfolio from ticker symbols and view its allocation
• Analyze a portfolio’s volatility, concentration, and sector breakdown
• Optimize an existing portfolio for maximum Sharpe ratio or minimum volatility
• Compare a user portfolio against a technology-sector example portfolio
• Explore available strategies and validate optimization settings

FAQ from MCP Portfolio Ideas

What is the guest key and how do I use it?

The README provides a guest key (mcp_portf_ideas_guest_…) that you set as the X-API-Key header in your MCP client configuration. It allows full access to all tools and strategies but operates in a shared workspace that is cleaned nightly.

What strategies are available at the guest level?

Guest users can use: maximize_sharpe_ratio, minimize_volatility, maximize_return, hrp, herc, equal_risk_contribution, maximum_diversification, and aliases like mean-variance. Additional strategies (target-return, CVaR, Black-Litterman) require a registered user key.

Are there any limitations on the assets I can use?

Bond ETFs (e.g., BND, AGG, TLT) are not available and will be excluded from portfolios. The service uses equity data from yfinance; calculations are based on historical data and scaled to a $100,000 account value for illustration.

What transport and authentication does the server use?

The server uses SSE transport at /mcp/sse/. Authentication is via a custom HTTP header X-API-Key. Any MCP client that supports remote SSE and custom headers can use this server.

Does the server provide financial advice?

No. The README explicitly states this service is for informational and educational purposes only, does not constitute financial advice, and past performance does not guarantee future results. Users should consult a qualified financial advisor.