For AI & agent builders

Your agent shouldn't improvise
compound interest

LLMs are excellent at conversation and terrible at multi-year financial arithmetic — and your users can't tell the difference until it costs them. Give your agent verified tools instead: it discovers the catalog, reads the contract, runs the model, and cites the spec version and assumptions in its answer.

🧰

Three tools, not thirty

list_models → get_model_contract → run_model. The contract tells your agent exactly what a model needs and what its outputs mean — tool use becomes self-correcting, not trial-and-error.

🧾

Answers that show their work

Every run returns its spec version, assumptions, triggered sentinels, and the sourced constants it used. 'Per the relocation model v1.0.0, break-even is 16 months' — an answer with receipts.

🔗

A handoff your users can touch

After the math, hand the user a live calculator with their numbers already loaded (/scenarios/relocation?currentSavings=40000). The conversation ends in a tool, not a wall of text.

One MCP endpoint, no auth

Streamable HTTP, in the official MCP registry as com.worthune/models. Add it to Claude or ChatGPT today, or call it from any MCP client library.

// MCP client config
{
  "mcpServers": {
    "worthune-models": {
      "url": "https://worthune.com/api/mcp/mcp"
    }
  }
}

// The agent then calls:
// list_models()
// get_model_contract({ model: "fire" })
// run_model({ model: "fire", inputs: { ... } })
Every model: published spec · independent second implementation, 250 verified cases each, green in CI · IRS/SSA constants with primary sources · zero silent changes. Explore the full catalog or read the docs.