Wardrobe Connect

Official2 tools

by GoodTurnStudio · JavaScript

Find clothes that match a description or a picture on eBay, new and pre-owned.

WardrobeConnect MCP Server (io.github.GoodTurnStudio/wardrobeconnect)

The WardrobeConnect MCP server provides a workflow to find clothes that match a user’s description or a picture on eBay, covering both new and pre-owned items. It is identified by the slug io-github-goodturnstudio-wardrobeconnect and the server name io.github.GoodTurnStudio/wardrobeconnect.

🛠️ Key Features

Clothing search using a text description
Clothing search using a picture
Availability includes new and pre-owned items on eBay

🚀 Use Cases

Locate eBay clothing that fits a textual requirement
Find similar or matching clothing from an uploaded image
Browse and filter both newly listed and previously owned items

⚡ Developer Benefits

Integrates image- and text-based clothing discovery for eBay listings
Provides a clear scope centered on wardrobe-related queries

⚠️ Limitations

Based on description alone: functionality details beyond eBay clothing matching are not specified

Captured live from the server via tools/list.

find_clothes

Find clothes that match a description. For when the user describes a piece of clothing they want ("black pleated wide-leg trousers", "cream cable-knit jumper like the one I saved"). Natural requests are fine ("a navy midi wrap dress for a summer wedding under £60 in a size 12"): the occasion is kept out of the search and returned in `occasion`, and a price or size in the words is used when max_price or size is not passed. Returns eBay listings in the Clothing, Shoes and Accessories category, closest matches first and then cheapest first (from a best-match and a price-sorted search), plus search links for other shops and a Facebook Marketplace search phrase. Listings must match the colour and fabric named, the same kind of garment (by title and by eBay category, so socks never answer a search for trainers), and men's or women's if given; kids' items only when asked for. Adult, fetish and costume listings are never shown, repeat listings of the same item count once, and listings far below the typical price for the search (usually bait or junk) are left out.

Parameters5

q	string	required	A clear description of the item: colour, type, fabric or style, brand if known.
size	string	optional	The user's size, for example M, XL, 12, UK 9, US 8.5, EU 39 or W32. Numbers without UK, US or EU are read in the country's usual sizes. Dress and shoe sizes are converted between UK, US and EU, and dress sizes to S, M and L. Listings that state a different size are left out; listings that state the size come first (size_listed true), then ones that don't say.
max_price	number	optional	Most the user wants to pay, in their local currency, shipping included.
condition	string	optional	Condition: new, used (pre-owned) or any. Default any. pre-owned and second-hand are read as used.
country	string	optional	Two-letter country code. Defaults to where the request comes from. Supported: US, GB, IE, CA, AU, DE, FR, IT, ES.

Raw schema

{
  "type": "object",
  "properties": {
    "q": {
      "type": "string",
      "description": "A clear description of the item: colour, type, fabric or style, brand if known."
    },
    "size": {
      "type": "string",
      "description": "The user's size, for example M, XL, 12, UK 9, US 8.5, EU 39 or W32. Numbers without UK, US or EU are read in the country's usual sizes. Dress and shoe sizes are converted between UK, US and EU, and dress sizes to S, M and L. Listings that state a different size are left out; listings that state the size come first (size_listed true), then ones that don't say."
    },
    "max_price": {
      "type": "number",
      "description": "Most the user wants to pay, in their local currency, shipping included."
    },
    "condition": {
      "type": "string",
      "enum": [
        "any",
        "new",
        "used"
      ],
      "description": "Condition: new, used (pre-owned) or any. Default any. pre-owned and second-hand are read as used."
    },
    "country": {
      "type": "string",
      "description": "Two-letter country code. Defaults to where the request comes from. Supported: US, GB, IE, CA, AU, DE, FR, IT, ES."
    }
  },
  "required": [
    "q"
  ],
  "additionalProperties": false
}

find_clothes_by_image

Find clothes that look like a picture. For when the user shares or points at a picture of an outfit or item (a Pinterest pin, a saved post, a photo). Takes the image's direct https link. Returns visually similar eBay listings from the Clothing, Shoes and Accessories category only, in eBay's look-alike order. If the picture does not look like one piece of clothing (the results are scattered across unrelated kinds), `matches` is empty and the say line asks what the item is. A short description in q makes results match it, which removes look-alikes of the wrong kind.

Parameters6

image_url	string	required	Direct https link to a JPEG, PNG or WebP picture, 6 MB at most.
q	string	optional	Which item in the picture, for example "the camel coat". Results must match it, and it is used as a fallback text search.
size	string	optional	Clothing or shoe size to match, for example M, 10 or UK 8.
max_price	number	optional	Highest price, in the local currency.
condition	string	optional	Condition: any, new or used.
country	string	optional	Two-letter country code for the eBay site and currency, for example US or GB.

Raw schema

{
  "type": "object",
  "properties": {
    "image_url": {
      "type": "string",
      "description": "Direct https link to a JPEG, PNG or WebP picture, 6 MB at most."
    },
    "q": {
      "type": "string",
      "description": "Which item in the picture, for example \"the camel coat\". Results must match it, and it is used as a fallback text search."
    },
    "size": {
      "type": "string",
      "description": "Clothing or shoe size to match, for example M, 10 or UK 8."
    },
    "max_price": {
      "type": "number",
      "description": "Highest price, in the local currency."
    },
    "condition": {
      "type": "string",
      "enum": [
        "any",
        "new",
        "used"
      ],
      "description": "Condition: any, new or used."
    },
    "country": {
      "type": "string",
      "description": "Two-letter country code for the eBay site and currency, for example US or GB."
    }
  },
  "required": [
    "image_url"
  ],
  "additionalProperties": false
}

README.md4 KBView on GitHub

goodturn-mcp

Remote MCP servers for 25 free, read-only everyday helpers by Good Turn Studio: money, health, jobs, travel, shopping, halal checks, football and scripture reading. One Cloudflare Worker turns each product's public OpenAPI spec into an MCP server, so any assistant that speaks MCP (Claude, ChatGPT, Alexa+, Cursor, VS Code and others) can use them.

One server per product: https://goodturn-mcp.pages.dev/<slug>/mcp
All products in one: https://goodturn-mcp.pages.dev/all/mcp (tool names start with the product slug)
Streamable HTTP, stateless JSON, MCP 2025-11-25 (and 2025-06-18, 2025-03-26), no sign-in
Every tool is read-only, with a title and annotations (readOnlyHint: true, destructiveHint: false, idempotentHint: true)
Nothing about the user is stored. Privacy: https://goodturn-mcp.pages.dev/privacy

Products

Product	Slug	What it does
Mizan	`mizan`	Shariah screening of US stocks under four standards, halal alternatives, purification, zakat
Halal or Not	`halalornot`	Halal checks for food, ingredients, medicines, stocks, funds and crypto
Fare Deals	`faredeals`	Cheap flights from any city, and the cheapest days to fly
Cheapest Price	`cheapestprice`	The lowest current price for a product, checked live
Spin My Day	`spinmyday`	A whole day out in any city, matched to weather, mood and diet
Cancel My Sub	`cancelmysub`	How to cancel US and UK subscriptions, and what a card charge is
Lower My Bill	`lowermybill`	Call scripts and low-cost plans to lower household bills
Price Drop Back	`pricedropback`	Store price-adjustment rules, deadlines and a claim message
Find My Money	`findmymoney`	Official unclaimed money searches in the US and UK
Supplement Check	`supplementcheck`	Supplement labels and doses against official upper limits
Plate Pal	`platepal`	Calories and nutrition for foods, chain meals and barcodes
Plan My Workout	`planmyworkout`	Workout plans and how to do any exercise
Med Bill Check	`medbillcheck`	US medical bills against Medicare rates, with dispute letters
Appeal My Claim	`appealmyclaim`	US health insurance denials, rights, deadlines and appeal letters
Job Spotter	`jobspotter`	Current US and UK job listings
Sync My Cycle	`syncmycycle`	Likely cycle phase today, with food and workout ideas
Wardrobe Connect	`wardrobeconnect`	Clothes that match a description or a picture
Form Guide	`formguide`	Football form, fixtures, results and tables
Books Like This	`bookslikethis`	Books like one you enjoyed, and screen adaptations
Talmud Scroller	`talmud`	Daily Talmud with Daf Yomi, and tractates in English
Scroll the Gurbani	`gurbani`	Daily Gurbani with English, any Ang or shabad
Gita Scroller	`gita`	Daily Bhagavad Gita verses with English
Scroll the Lore	`lore`	Greek and Norse myth, Homer, Hesiod and the Poetic Edda
Stoic Scroller	`stoics`	Marcus Aurelius, Epictetus and Seneca
Bible Swipe	`bible`	Daily Bible readings and any passage

How it works

build.py reads each product's live openapi.json and writes products.js: one tool per GET operation, named from its operationId, with inputs from its parameters. Descriptions are tidied into plain tool descriptions.
worker.mjs is the MCP server: initialize, tools/list and tools/call, which calls the product's own API and returns its answer. Answers with a say field carry a ready-made summary.
servers.json lists every server for the official MCP Registry (as io.github.GoodTurnStudio/<slug>). The Publish to the MCP Registry workflow logs in with GitHub OIDC and runs publish.py; publish.ps1 does the same from Windows.

Build and test

code

npm install
npm run build          # python3 build.py, then bundle to deploy/_worker.js
npm test               # every product through the worker, against the live APIs
npm run test:live      # every tool through the deployed servers with the MCP SDK client

Deploy deploy/ as a Cloudflare Pages project in advanced mode (_worker.js).

Licence

MIT for this code. Each product's API, data and answers belong to that product and are served under its own terms, linked from its page at https://goodturn-mcp.pages.dev/<slug>.

Install

Tools (2)

find_clothesfind_clothes_by_image

Remote endpoint

Streamable HTTP

Hosted server - connect over the network, no local install.

url

https://goodturn-mcp.pages.dev/wardrobeconnect/mcp

Add to VS Code Add to Cursor

claude_desktop_config.json

json

{
  "mcpServers": {
    "wardrobeconnect": {
      "command": "npx",
      "args": [
        "-y",
        "mcp-remote",
        "https://goodturn-mcp.pages.dev/wardrobeconnect/mcp"
      ]
    }
  }
}

Desktop config is stdio-only; this bridges via mcp-remote. Native remote: Settings > Connectors.