Not a toy menu. A build map for deterministic finance.

Live does not mean reckless. Every model computes the layer it can defend and exposes the boundary when a banker-looking number would require unsupported assumptions.
Model-implied valuation workbook from EDGAR facts, sealed formulas, AURORA refusals, and LUNA chain.
JL_DCF.xlsxCFI_DCF.xlsxSaaS_DCF.xlsxDCF_Valuation.xlsxComparable-company analysis with SIC peer resolution, fiscal alignment, sealed multiple definitions, and peer-quality review.
multiples_definitions.yamlpeer_resolver.yamlgaap_canonical_tags.yamlPrecedent transaction analysis from EDGAR deal filings, point-in-time terms, unaffected-price checks, and implied valuation ranges.
precedent_resolver.yamldeal_terms_taxonomy.yamldeal_filing_cascade.yamlLinked operating spine from EDGAR-backed DCF facts: revenue, EBITDA, EBIT, NOPAT, capex, working capital, and FCF.
Stmt3.xlsxStmt3_Compact.xlsxStmt3_Full.xlsxBase, bear, and bull cases with explicit assumption deltas and sensitivity tables. Built to prevent hidden assumption drift.
Scenarios.xlsxFPA_Model.xlsxAcquirer/target validation, sources and uses, purchase-price bridge, synergy assumptions, and explicit EPS refusal boundary.
MA_Model.xlsxMA_Accretion.xlsxMA_Intro.xlsxSponsor-return model with editable leverage, FCF sweep, exit multiple, debt paydown schedule, MOIC, and IRR.
Energy_LBO.xlsx01_cipher_lbo.mdDebt capacity screen with leverage, coverage, cash-flow cushion, deterministic flags, and covenant interpretation refusal boundary.
02_cipher_credit_PUNTED.md15_cipher_covenant.mdSingle-segment SOTP shell with editable core EV weight and non-core asset adjustment; refuses fake segment allocation.
RE_Modeling.xlsxRE_CapRate.xlsxInsurance_Allstate.xlsxOne-page public-company profile combining EDGAR facts, DCF, COMPS, market data, and LUNA links.
DCF + COMPS + market dataArchitecture comparison frame plus paste-and-score LLM challenge for ALCHEMIST model outputs.
score_llm.pybuild_llm_prompts.pyALCHEMIST challenge rubricStart with DCF and COMPS to prove the data spine, then jump to Scenarios, LBO, Credit, and Benchmark. The benchmark page is where the LLM comparison lands cleanly: same task, same input, different architecture.