# PricingLab

> PricingLab by Raganel Consulting, powered by MatLogica AADC. Priced financial and insurance models that return the value and every sensitivity from one adjoint pass, with a built-in check that recomputes the sensitivities by bump-and-revalue on the same scenarios. Use it from any MCP client, from Excel, or as a local install. Operated by Raganel Consulting Limited EOOD (Bulgaria). Contact: info@mypricinglab.com

Every number in an answer should come from a tool result. Start with `list_models`, then `describe_model` for inputs, units and defaults. Any input left out takes its default.

## Models

- [Variable annuity: lifetime withdrawal guarantee](https://mypricinglab.com/models/variable_annuity.md): value of the insurer's guarantee on a variable annuity (GMWB with annual ratchet and roll-up, mortality and dynamic lapse), with delta, vega and rho; plus a nested (scenarios within scenarios) run of the liability and hedge ratio over 10 years.
- [Insurance ALM: Solvency II balance sheet](https://mypricinglab.com/models/insurance_alm.md): best-estimate liability, risk margin, own funds, SCR by module and solvency ratio for a simulated 1,200-member bulk-annuity book backed by 21 bonds, with the sensitivity of the liability, assets and own funds to all 21 assumptions.
- [Gas storage: facility value under an optimised operating policy](https://mypricinglab.com/models/gas_storage.md): value of a natural-gas storage facility run by an optimised daily injection and withdrawal policy under a two-factor price model, with sensitivities to all twelve monthly forwards and the interest rate, and the intrinsic value on today's forward curve alongside. On the hosted service the facility and curve shape are fixed; the local install trains a policy for any facility.
- [Barrier option: knock-out and knock-in](https://mypricinglab.com/models/barrier_option.md): price of a continuously monitored barrier option (down or up, out or in, call or put) under Black-Scholes, with delta, vega and rho; Brownian-bridge Monte Carlo checked against the closed-form prices.
- [Asian option: arithmetic average](https://mypricinglab.com/models/asian_option.md): price of an arithmetic-average Asian call under Black-Scholes, with delta, vega and rho; price checked against QuantLib's Monte Carlo engine.
- [Basket option: two assets](https://mypricinglab.com/models/basket_option.md): price of a call on an equally weighted two-asset basket under correlated Black-Scholes, with both deltas, vega and rho; price checked against QuantLib's Monte Carlo basket engine.

## Tools

- `list_models`: list the available models with a one-line summary each.
- `describe_model`: inputs (name, unit, default, range, hosted-service cap), outputs and extra functions of one model.
- `price_model`: value and all sensitivities from one adjoint pass; `inputs` overrides defaults by name.
- `check_against_bumping`: recompute the sensitivities by central bump-and-revalue on the same scenarios; reports each gap, whether they agree within tolerance, and the time each method took.
- `run_model_function`: run a model's extra calculation, for example `model="variable_annuity"`, `function="nested"`.

## Ways to connect

- [Hosted MCP endpoint](https://mypricinglab.com/lab/mcp): MCP over Streamable HTTP, no authentication, free, made-up trades with capped sizes. Add it as a remote server in any MCP client.
- [Excel add-in and workbooks](https://mypricinglab.com/excel): install the PricingLab add-in from its manifest and open the demo workbook for a model.
- [Local install](https://mypricinglab.com/install): sent on request (Linux x86-64 today) from info@mypricinglab.com; then `pricinglab mcp` (MCP over stdio) or `pricinglab serve` (local web server for the Excel workbooks). No size caps.

## Contact

- [Email info@mypricinglab.com](mailto:info@mypricinglab.com): a private hosted service with login, or running the models on your own book inside your own network.