Optionsahoy MCP

What AlvisoOculus/optionsahoy-mcp MCP server does

AlvisoOculus/optionsahoy-mcp MCP server exposes equity-compensation and tax calculations as Model Context Protocol tools. It is intended for structured scenarios involving stock options, restricted stock units, concentrated positions, hedges, funding goals, and qualified small business stock. The calculations cover federal ordinary income, long-term capital gains, AMT, FICA, and NIIT considerations, together with state treatment across all 50 states and DC. State AMT modeling is specifically described for California, Colorado, Connecticut, and Minnesota.

The server is a calculation service, not a general tax-advice conversation tool. Its inputs describe a financial scenario, and its responses provide computed comparisons, schedules, prices, or qualification results.

How it works

AlvisoOculus/optionsahoy-mcp MCP server is available at https://optionsahoy.com/mcp as a remote HTTP MCP endpoint. MCP-compatible clients can register that URL and call the exposed tools without supplying authentication or creating an account. The repository also documents a local npm stdio option using npx -y optionsahoy-mcp.

The ISO optimizer searches a discretized multi-year candidate space and refines the result at the share level. Other planners use deterministic searches that account for tax brackets, timing, and stated objectives. The response from the API is intended to correspond to the figures produced by the related OptionsAhoy browser calculators.

Tools and capabilities

The documented tools are:

• amt_iso_optimize: Finds a multi-year ISO exercise schedule while modeling AMT credit recovery, grant expiration, and the post-termination exercise period.
• nso_calculate: Compares selling NSOs at exercise with holding for potential long-term capital gains, including federal, state, and FICA effects.
• rsu_sell_vs_hold: Compares selling RSUs at vest with holding them, including withholding and marginal-rate differences.
• concentration_analyze: Evaluates downside exposure and compares selling, holding, and hedging a concentrated single-stock position.
• protective_put_price: Prices protective puts, zero-cost collars, and put spreads using Black-Scholes inputs.
• qsbs_check: Tests Section 1202 qualification across six statutory requirements and reports applicable exclusion and state-conformity results.
• equity_funding_plan: Builds multi-year sale plans for reaching an after-tax funding target by a deadline and returns named plans plus a risk/wealth frontier.
• rsu_lot_optimize: Selects RSU lots and sale dates to reduce computed tax while considering long-term holding periods, loss carryforwards, and bracket spreading.

Setup and configuration

For a hosted setup, add https://optionsahoy.com/mcp as a remote HTTP server in the MCP client. Claude Desktop also has a downloadable optionsahoy.mcpb package in the repository's latest releases. A local npm stdio invocation is documented separately from the hosted endpoint.

AlvisoOculus/optionsahoy-mcp MCP server does not require an API key, token, account, or listed environment variable for the hosted endpoint. Scenario inputs should be supplied through tool arguments, such as share counts, strike price, fair-market value, filing status, income, state, dates, growth assumptions, or volatility assumptions where applicable.

Limitations and notes

The outputs depend on the assumptions and inputs provided to each calculation. They should not be treated as a substitute for individualized advice from a tax professional or financial adviser. The server models the tax and pricing rules described in its documentation; it does not indicate that a particular transaction is suitable for a user.

AlvisoOculus/optionsahoy-mcp MCP server focuses on deterministic computation. It does not claim to infer missing facts reliably, and invalid requests are rejected with field-level validation errors according to the published schema. The repository reports automated testing and independent comparison cases, but users should still review assumptions before acting on a result.