Command Palette

Search for a command to run...

UnylyUnyly
Весь каталог

Resonant Grocery Flywheel

БесплатноНе проверен

Local-first, human-approves-everything grocery replenishment as a ResonantOS add-on: runway analysis, fail-closed dietary checks, and a next-cart draft that str

GitHubEmbed

Описание

Local-first, human-approves-everything grocery replenishment as a ResonantOS add-on: runway analysis, fail-closed dietary checks, and a next-cart draft that structurally cannot check out. No accounts, no cloud, no telemetry — nothing leaves your machine.

README

Local-first, human-approves-everything groceries. Those two values are the whole reason this tool belongs in this community: your food and habit data never leaves your machine, and the flywheel never spends a dollar you did not personally approve. It tracks what you bought, notices what you've used, flags what's going stale, and drafts the next cart — you say yes.

The engine is grocery-flywheel (MIT, vendored files sha256-pinned in vendor/VENDOR-MANIFEST.json), vendored byte-identical under vendor/ and wrapped by a thin local service. The wrapper adds no dependencies: Python 3.10+ standard library only. No accounts, no cloud, no telemetry.

Honest surface

This add-on wraps exactly the verbs the upstream MCP server exposes to AI assistants — the tool's real headless surface — plus its status:

Verb Upstream tool What it actually does
groceryflywheel.status — version, busy flag, last dashboard
groceryflywheel.analyze analyze_replenishment_state runway, depletion, roles, freshness, dietary evaluations, substitutions
groceryflywheel.dietary evaluate_dietary fail-closed safety: restrictions without current label evidence come back needs_review, never safe
groceryflywheel.sourcing summarize_sourcing_research open sourcing questions + decision boundaries recorded in the state
groceryflywheel.plan plan_next_cart next-cart draft; every item is needs_human_approval, and the vendored engine itself raises unless checkout_available is false
groceryflywheel.dashboard render_replenishment_dashboard the dark command-center HTML, written only under var/dashboards/

What is deliberately NOT wrapped, declared plainly:

  • Checkout. It does not exist anywhere in the upstream engine by design (checkout_available is structurally false; upstream ships a test that fails if that ever changes). Nothing was removed — there is nothing to remove.
  • state_path / output_path. The upstream MCP tools accept filesystem paths; this wrapper does not. State arrives inline as state_json (at most 60000 characters — it must fit the 64KB request body), and dashboard HTML is written only inside this addon's var/dashboards/. No filesystem capability is requested; the manifest requests zero capabilities.
  • Corrections, capture-visit, and retailer-import CLI flows. Those are upstream CLI workflows (subprocess entry points), not part of the MCP surface; this service spawns no subprocesses, so they are out of scope here.
  • Upstream quirk, preserved and declared: upstream's analyze_replenishment_state schema advertises objective but its handler does not apply it (only evaluate_dietary and plan_next_cart do). This wrapper mirrors upstream exactly; to get an objective-aware analysis, put objective in the state document itself.

Privacy posture

Food and habit data is personal. Following the Innerscape posture: the data surface is data-bearing, and nothing leaves the machine — the service binds loopback only, calls the vendored engine in-process, writes only under var/, requests zero capabilities, and keeps no telemetry. Every response and every persisted file passes through home-path redaction before it exists. The only state used by the test suite is the upstream sanitized example and synthetic fixtures built in the tests.

Running it

python3 server.py          # listens on http://127.0.0.1:4899 (the manifest entrypoint)

curl -s http://127.0.0.1:4899/health
curl -s -X POST http://127.0.0.1:4899/ -H 'Content-Type: application/json' \
  -d '{"method":"groceryflywheel.plan","params":{"state_json":"...","mode":"pickup"}}'
curl -s -X POST http://127.0.0.1:4899/ -H 'Content-Type: application/json' \
  -d '{"method":"groceryflywheel.dashboard","params":{"state_json":"...","filename":"week.html"}}'

A ready-made state to try lives at vendor/examples/sample_state.json (upstream's sanitized example). Environment: GROCERYFLYWHEEL_PORT (dev only — the manifest declares 4899). The service binds 127.0.0.1 only, spawns no subprocesses, and exits with code 78 if the port cannot be bound.

Tests

python3 -m unittest discover -s tests        # wrapper suite
sh run-validator-check.sh <path-to-2.0.0-alpha-clone>  # manifest vs the real validator
cd vendor && PYTHONPATH=. python3 -m pytest tests/ -q   # upstream suite, unmodified (188 tests)

vendor/ is hash-pinned in vendor/VENDOR-MANIFEST.json: the wrapper suite fails loudly if any vendored engine file, vendored upstream test, or vendored example drifts by a single byte. The vendored upstream suite is upstream's own tests unmodified, minus test_quickstart.py (it shells out to installed grocery-flywheel console scripts and a docs/ tree — environmental, not engine, coverage; the wrapper does not expose the CLI).

License

MIT — see LICENSE. The vendored grocery-flywheel engine is MIT, Copyright (c) 2026 Kyanite Labs (Simon Gonzalez de Cruz) — see the upstream repository.

from github.com/simongonzalezdc/resonant-grocery-flywheel

Установка Resonant Grocery Flywheel

У этого сервера нет опубликованного пакета — он собирается из исходников. Открой репозиторий и следуй инструкции в README.

▸ github.com/simongonzalezdc/resonant-grocery-flywheel

FAQ

Resonant Grocery Flywheel MCP бесплатный?

Да, Resonant Grocery Flywheel MCP бесплатный — установка в пару кликов через Unyly без оплаты.

Нужен ли API-ключ для Resonant Grocery Flywheel?

Нет, Resonant Grocery Flywheel работает без API-ключей и переменных окружения.

Resonant Grocery Flywheel — hosted или self-hosted?

Self-hosted: сервер запускается локально на твоей машине командой из раздела установки.

Как установить Resonant Grocery Flywheel в Claude Desktop, Claude Code или Cursor?

Открой Resonant Grocery Flywheel на unyly.org, выбери вкладку своего клиента (Claude Desktop, Claude Code, Cursor) и нажми Install — конфиг сгенерируется автоматически, без правки JSON.

Похожие MCP

Opencode Omniroute Plugin

OpenCode plugin for the OmniRoute AI Gateway. Drives dynamic model discovery, /connect auth flow, and multi-instance OmniRoute providers via the official @openc

GitHub Actionsавтор: GitHub Actions
91

Roblox Studio

Enables AI coding tools to control Roblox Studio for workspace exploration, instance manipulation, and script management. It provides tools for playtesting, sce

paralovавтор: paralov

Antigravity Drawio

MCP server that enables AI assistants to create, parse, render, and validate Draw.io diagrams programmatically.

S-SUJAN-Sавтор: S-SUJAN-S

Fetch

Web content fetching and conversion for efficient LLM usage.

автор: Community

AWS KB Retrieval

Retrieval from AWS Knowledge Base using Bedrock Agent Runtime.

modelcontextprotocolавтор: modelcontextprotocol

Spring AI MCP Server

Provides auto-configuration for setting up an MCP server in Spring Boot applications.

автор: Community

llm-analysis-assistant

A very streamlined mcp client that supports calling and monitoring stdio/sse/streamableHttp, and can also view request responses through the /logs page. It also

xuzexin-hzавтор: xuzexin-hz

MCP-Agent

A simple, composable framework to build agents using Model Context Protocol by [LastMile AI](https://www.lastmileai.dev)

lastmile-aiавтор: lastmile-ai

Spring AI MCP Client

Provides auto-configuration for MCP client functionality in Spring Boot applications.

автор: Community

mcp.natoma.ai

A Hosted MCP Platform to discover, install, manage and deploy MCP servers by [Natoma Labs](https://www.natoma.ai)

автор: Community

Compare Resonant Grocery Flywheel with

Не уверен что выбрать?

Найди свой стек за 60 секунд

Автор?

Embed-бейдж для README

Похожее

Все в категории ai