Use POP from ChatGPT, Claude or any AI assistant

Claude (custom connector)

1. In Claude, open Settings → Connectors and choose Add custom connector.
2. Name it Portfolio Optimizer Pro and paste the URL https://portfoliooptimizerpro.com/mcp. Leave OAuth settings empty.
3. Select Add. In a chat, turn the connector on from the tools menu, then ask one of the prompts below.

ChatGPT (developer mode connector)

1. Open Settings → Apps & Connectors → Advanced settings and turn on Developer mode (available on plans that support it).
2. Back in Apps & Connectors, choose Create. Name it Portfolio Optimizer Pro, paste https://portfoliooptimizerpro.com/mcp as the MCP server URL and pick No authentication.
3. Confirm and create it. In a new chat, add it from the + menu under Developer mode, then ask one of the prompts below.

Menu names in assistant apps change from time to time; look for “connectors” or “MCP”.

Other assistants and developers

Any MCP client that supports streamable HTTP can use https://portfoliooptimizerpro.com/mcp. The same tools are a plain JSON API at https://portfoliooptimizerpro.com/api/public/agent/{tool}, described in the OpenAPI 3.1 spec. A summary for AI crawlers is in llms.txt.

Example prompts

• “Use Portfolio Optimizer Pro to grade my portfolio: 60% VTI, 30% VXUS, 10% BND.”
• “What held the grade down for 100% VTI, and how does the grading work?”
• “What do my funds cost a year? VOO 50%, QQQ 30%, SCHD 20%.”
• “Do VTI, VOO and QQQ overlap?”

Tools

grade_portfolio(holdings: [{ticker, weight}])

A+ to F structural grade, 0–100 score, any ceiling applied, each factor's sub-score and measured fact, top findings, data dates and a link to the same graded checkup on the site.

fund_fee_cost(holdings: [{ticker, weight}])

Weighted expense ratio from dated issuer records and the fee cost over 10 and 30 years on $100K, same formula as the expense ratio calculator.

check_overlap(holdings: [{ticker, weight}])

Known index overlap and related fund mandates, labelled as category-based rather than measured shared holdings.

explain_grading(none)

The factors, scoring rules, caps, thresholds, letter bands and data source behind the grade.

How the grade is calculated: methodology.

Privacy

No account or login. Holdings you send are processed only to compute the result and are not stored. We keep daily counts of tool calls by tool and the app name the assistant reports, and of new price lookups, never holdings, names or emails; call limits use a one-way keyed hash of the network address, deleted after about two days. Links back to this site carry holdings after the “#” in the address, which browsers never send to servers; the page clears it on arrival. See the privacy policy.