Who it is for
Recipe developers and food product teams reasoning about ingredient pairings and substitutions.
AI Connectors · Data & Analytics
Computational flavour pairing and substitution across 1,790 ingredients
Overview
Epicure is a flavour model rather than a recipe database: 1,790 ingredients placed on sensory and cultural axes, with pairing scores computed from bundled model artefacts. KAIKAKU.AI states the results are deterministic and that no external model is called, so the same query returns the same answer each time.
The connector exposes that model read-only. Ingredients can be compared on a named axis, scored as pairs, or located on a flavour atlas; substitutions come from nearest-neighbour lookups. Documentation is published and the connector is community-listed, so it is self-published rather than reviewed.
Recipe developers and food product teams reasoning about ingredient pairings and substitutions.
Thirteen tools answered without credentials. They split into scoring (pairing_score, compare_on_axis, flavour_correlations), navigation (find_pairings, neighbors, pareto_navigate, where_on_atlas), and model introspection (list_targets, list_factors, ingredient_on_factor), with morph walking the path between two ingredients.
Availability
Epicure is listed in the Claude connectors directory only. KAIKAKU.AI has no app under the same domain in the ChatGPT directory; 392 of the 3,150 products here are on both marketplaces, and this is not one of them.
The MCP server behind a Claude connector is usually the same server a ChatGPT app would need, so the absence is a distribution decision rather than a technical one.
Claude listing
Epicure publishes 13 tools to Claude. That is a mid-sized surface: 54.3% of the 1,253 Claude-listed connectors expose fewer.
The median Claude-listed connector filed under Data & Analytics publishes 12, so this one runs one tool above its bucket's median and ranks 84th of 198 by tool count.
It falls in the 11-20 tools band, which holds 22.5% of the Claude directory.
A direct call to the endpoint returned 13 tools, matching the 13 the listing publishes. Probed 2026-08-07.
compare_on_axispairing_scorefind_pairingsflavour_correlationscultural_profileneighborsmorphlist_targetslist_factorsingredient_on_factorpareto_navigateclosest_modewhere_on_atlasEpicure publishes a remote MCP endpoint. Claude connects to it over HTTP; there is nothing to install locally.
Every connector in the Claude directory is remote, all 1,253 of them, so transport is not a differentiator there.
https://epicure-mcp.kaikaku.ai/mcp Epicure is community-listed in the Claude directory, the tier 770 of 1,253 connectors sit in (61.5%). It is the default for a self-published server, and says nothing either way about quality.
In the publisher's words
Epicure gives Claude 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. Results are deterministic and computed from bundled model artefacts; Epicure does not call an external model.
Quoted from the connector's own Claude directory listing as captured in August 2026. Node8 did not write it and does not vouch for it.
Nearby
Agent-driven research on people and companies
ChatGPT
Exact Online finance data for fractional CFOs and founders
Claude
Structured sentiment and topic data from iOS and Android reviews
Claude
Cited research answers over a curated enterprise source set
Claude
Company headcount, hiring, and attrition analytics for market research
Claude
Location and proximity questions answered from spatial data
ChatGPT
This is an independent catalogue entry compiled by Node8 from Epicure's public marketplace listing as it stood in August 2026. Listing facts come from the marketplaces; the comparisons against the rest of the directory and any live-endpoint measurements are Node8's. Node8 is not affiliated with KAIKAKU.AI, and inclusion here is not an endorsement.
Node8 builds MCP servers end to end: deciding which tools are worth exposing, the authentication and permission model, the review submissions, and the onboarding that gets them used. One server serves ChatGPT, Claude, and Microsoft Copilot.