build_myga_ladder_scenario
ChatGPTBuild a transparent equal-allocation educational MYGA ladder for 2–5 requested terms using current public rate rows. Shows candidates, assumptions, separate maturity dates, citations, and unavailable rungs; it is not personalized advice.
compare_annuity_products
ChatGPTCompare 2–4 exact canonical ARHQ product URLs under one shared state/premium/age/timeline scenario. Returns factual fields and per-source citations without ranking suitability or recommending a purchase.
get_annuity_carrier
ChatGPTGet one public ARHQ carrier profile by carrier slug, including dated public ratings and canonical product-page coverage. Returns unavailable rather than guessing when the slug does not resolve.
get_annuity_product
ChatGPTGet normalized public details for one canonical ARHQ MYGA/FIA product page or RILA review URL. For an exact mapped FIA and supplied state, also returns a product-pinned authoritative income illustration with annual/monthly income, income base, fees, provenance, and a bounded year ledger—or a precise unavailable reason. Never accepts or returns licensed source IDs; unavailable fields are null and source failures are explicit.
get_annuity_rate_changes
ChatGPTGet bounded public MYGA/FIA rate and availability changes observed by the daily CANNEX refresh. Includes prior/new values, observation time, cohort size, source context, and canonical ARHQ links.
get_annuity_rate_index
ChatGPTGet current values from the coherent 18-series public Annuity Rate Index contract. Returns normalized percentage points or monthly income, exact per-series cohort/sample sizes, source vintage, and explicit trend availability. Null delta and an empty series mean history is withheld—not no change.
get_arhq_data_health
ChatGPTUse when assessing freshness, outages, or whether zero results are trustworthy. Returns data.sources with bounded per-family status, source vintage, record count, coverage statement, canonical URL, and explicit failure reason; it does not expose infrastructure or licensed source details.
get_arhq_methodology
ChatGPTExplain ARHQ MCP provenance, numeric context, null semantics, response bounds, citation rules, and recommendation limits. Use this when evaluating how to cite or interpret another ARHQ tool result.
research_annuity_income_and_legacy
ChatGPTResearch source-backed fixed indexed annuity income-rider candidates for one exact issue age, future income-start age, premium, and U.S. state, with an optional public product/carrier query, optional exact comparisons at other income-start ages, and an optional first-year annual-income target. Query focus resolves only through public FIA mappings and curation, uses the cached authoritative product-pinned illustration DAL, and fails closed without unrelated generic candidates; query and comparison_income_start_ages cannot initially be combined. For each of at most five returned candidates, returns annual/monthly income, income base, rider fees, scenario/provenance, and a bounded year ledger—or a precise unavailable reason. Also returns explicit source-versus-derived monthly-income basis, factual tradeoff leaders, target arithmetic, timing sensitivity, compact shortlist cards, link-coverage metrics, legacy-evidence gaps, field-level rate quarantines, and a candidate-specific carrier-illustration verification packet. In the user-facing answer, render each details_link beside its product and use its source_citation within the shortlist card; disclose fia_research fallbacks as category links rather than full product pages or candidate-specific verification. Use when a user wants retirement income while preserving possible value for beneficiaries; this is not a personalized recommendation or suitability ranking.
search_annuity_products
ChatGPTSearch public ARHQ MYGA, FIA, RILA-review, and FIA-income records. Returns bounded, paginated results with canonical ARHQ URLs, explicit unavailable families, source vintage, and null-safe rate fields.
search_myga_rates
ChatGPTUse for current MYGA questions that specify a guarantee term. Returns data.results with rate_percent, rate_kind, carrier, AM Best rating, premium bounds, canonical product URL, exact applied filters, source vintage, and filter-bound pagination; factual sorting is not a recommendation.