I would like you to be able to access your Partec portfolio via an MCP server (Model Context Protocol) — i.e. a standardized interface so that you can directly access and work with your own portfolio data with models of your choice (e.g. various LLMS/agents). Why that makes sense: Especially when it comes to portfolios with a large number of individual stocks, the same question constantly arises: Should I buy back position X? Rebalances? Reduce risk? Thesis A vs. Thesis B? This can be played through extremely well with a structured model — but only if the model can “see” the portfolio clearly (allocation, entry prices, performance, sectors, countries, profits/losses, cash, watchlist, etc.). What the interface should provide (examples): • Read-only access (at least): Holdings, Weights, Income, Current Value, P&L, Asset Classes, Cash, Transactions • Optional: watchlist/ goals/ own notes per position • Optionally later: Prepare actions (e.g. “order suggestions” as a draft, not execute directly) MCP added value: MCP makes the whole thing compatible with many tools/clients and allows users to use their preferred model without Partec having to build “the one AI” themselves. Result: better decisions, faster portfolio reviews, less manual copy/paste.