build_tasting_flight
ChatGPTBuild a 2-6 product tasting flight from exact variant IDs. Orders and explains selections only from recorded facts; unknowns remain explicit. Flavor notes are sensory descriptors, not ingredients. Not for allergy-safety decisions.
catalog_status
ChatGPTRead the source, capture time, freshness state, product count, and exact-variant coverage of the Bar & Cocoa public storefront snapshot before interpreting recommendations.
compare_chocolates
ChatGPTCompare 2-4 distinct exact catalog products across freshly checked price and availability when available, cacao percentage, maker, cacao origin, documented flavor notes, and explicit unknown texture. Each exact variant is labeled current, stale, or unverified. Returns conditional, fact-grounded choice guidance without scoring quality or predicting taste. Fails if any ID is absent.
get_chocolate
ChatGPTRead one exact product by ID, check its exact current Shopify public price and availability, and, when available, hydrate bounded source-backed editorial facts from its exact public product page. Commerce status identifies current, stale, or unverified values; only current values may be described as current. Use an ID returned by search_chocolates.
get_chocolate_guide
ChatGPTRead source-reviewed maker profiles and editorially reviewed tasting lessons. Use this to explain maker location versus cacao origin, percentage, texture, controlled comparisons, or how to taste. Maker-level context never overrides an exact product record.
recommend_next_chocolates
ChatGPTRecommend different exact catalog products after a known chocolate. A bounded candidate pool is checked against exact current variant price and availability before stock or USD budget constraints are applied. Direction and budget are hard constraints. This is preference guidance, not a taste prediction, allergy-safety judgment, or delivery guarantee.
search_chocolates
ChatGPTUse this when someone wants to find Bar & Cocoa chocolate by budget, maker, cacao origin, cocoa percentage, or documented tasting notes. Searches a maintained public storefront snapshot with exact variants and strict filters. Page-backed labels discovered only when exact detail is checked are not indexed by broad search and may be cached for up to 24 hours. Do not use for delivery guarantees, allergy safety judgments, purchases, or a complete market comparison. Prices and availability reflect the capture time. Unknown metadata cannot satisfy a filter. Zero matches means no documented snapshot match, not verified absence.
search_gifts
ChatGPTUse this when someone wants a Bar & Cocoa gift by occasion, recipient, budget, or curated presentation. A bounded set of reviewed gift candidates and exact variants is checked against current Shopify public price and availability before stock or USD budget constraints are applied. Treat every filter as a hard constraint. Never promise delivery timing, allergy safety, gift-note availability, or packaging that the result does not explicitly support.
