Day 22 — MCP Integration
Your client calls. “How's my portfolio doing?” You open four tabs, three spreadsheets, and a PDF.
An MCP server that tracks client portfolios across accounts, custodians, and asset classes. Holdings, allocation drift, risk analysis, transaction history, and alerts — all from one question instead of four systems.
Example Queries
Tools
| Tool | What It Does |
|---|---|
| list_clients | List and search wealth management clients. Shows portfolio summary, risk profile, advisor assignment, and total AUM. |
| get_portfolio | Get a client's full portfolio breakdown across all accounts — holdings, allocation, performance, and drift analysis vs. model portfolio. |
| get_account | Get detailed view of a single investment account — holdings by weight, sector breakdown, recent transactions, and performance. |
| list_transactions | Search and filter transactions across accounts. Filter by client, type, or symbol. Shows totals by transaction type. |
| portfolio_analysis | Run risk and allocation analysis — diversification score, concentration risks, sector exposure, and rebalancing recommendations. |
| get_alerts | Get portfolio alerts — concentration warnings, rebalance triggers, RMD deadlines, contribution reminders, and performance flags. |
Client Book
Retirement + legacy · David Park
Capital preservation + income · David Park
Growth + accumulation · Sarah Mitchell
Trust mgmt + philanthropy · Sarah Mitchell
College fund + retirement · David Park
Growth + early retirement · Sarah Mitchell
Why This Build
Wealth advisors live in a world of fragmented data. Client portfolios are spread across custodians, account types, and asset classes. Answering “how am I doing?” means logging into Schwab, pulling a Fidelity report, checking the 529 balances, and doing mental math on allocation drift.
Most portfolio management software costs $10K+/year and still requires manual data entry. The advisor ends up spending more time pulling reports than actually advising clients. Meanwhile, the client just wants a straight answer.
This server manages six clients with $16M+ in combined assets across 14 accounts. It tracks 41 holdings, computes allocation drift against model portfolios, scores diversification, flags concentration risks, and surfaces alerts for RMDs, rebalancing, and contribution limits. Ask a question, get the answer.
Demo Data
6
Clients
14
Accounts
41
Holdings
20
Transactions
20
Model Allocations
8
Alerts
Six clients from conservative retirees to aggressive accumulators. Fourteen accounts spanning trusts, IRAs, Roth IRAs, 529 plans, SEP IRAs, and charitable remainder trusts. Four risk profiles with model allocations. Realistic edge cases — single-stock concentration risk, missing international exposure, RMD deadlines, and a client targeting early retirement at 50.
Setup
cd day22-wealth-management
npm install && npm run seed && npm run buildThen add the config to Claude Desktop and restart.
If your client book lives across spreadsheets, PDFs, and custodian dashboards — let's talk about bringing it all into one conversation.