Epicure

Computational flavour intelligence for culinary agents

Epicure gives your AI read-only access to a computational flavour model spanning 1,790 ingredients. Explore pairings, compare ingredients on sensory and cultural axes, discover substitutions and neighbours, locate ingredients on a flavour atlas, and navigate interpretable flavour transformations.

MCP connector

Server URL: https://epicure-mcp.kaikaku.ai/mcp
Login: None
Protocol: MCP 2025-11-25
Instructions: Read-only access to the Epicure ingredient-embedding model: 1,790 ingredients embedded as 300-dimensional vectors trained on a 4.14M-recipe multi-language corpus. The model captures both recipe co-occurrence (which ingredients appear together) and learned flavour structure (cuisine, sensory, nutrient, processing axes). Ingredient names are matched deterministically (free text like "fresh ginger" maps to a canonical name); no fuzzy LLM fallback.
Tool families:
Pairing exploration: , , neighbors.
Measurement: , , , .
Spatial context: , .
Directed transformation: morph and .
Valid target catalogues: and .
returns a diversified cluster-and-bridge graph for open-ended pairing and recipe-design questions. measures one exact pair, while neighbors returns direct single-seed similarity. morph represents an explicit directional transformation rather than undirected exploration.

Tools

Reads data13

Skills

None in the store listings or on the server

MCP connector

Server URL

https://epicure-mcp.kaikaku.ai/mcp

None

Protocol

MCP 2025-11-25

Instructions

Read-only access to the Epicure ingredient-embedding model: 1,790 ingredients embedded as 300-dimensional vectors trained on a 4.14M-recipe multi-language corpus. The model captures both recipe co-occurrence (which ingredients appear together) and learned flavour structure (cuisine, sensory, nutrient, processing axes). Ingredient names are matched deterministically (free text like "fresh ginger" maps to a canonical name); no fuzzy LLM fallback.

Tool families:

Pairing exploration: , , neighbors.

Measurement: , , , .

Spatial context: , .

Directed transformation: morph and .

Valid target catalogues: and .

returns a diversified cluster-and-bridge graph for open-ended pairing and recipe-design questions. measures one exact pair, while neighbors returns direct single-seed similarity. morph represents an explicit directional transformation rather than undirected exploration.