Agent Academy

Agent Academy is a free, source-linked library of skills, prompts, MCP servers and bot templates, organised into courses by profession, for people who hand real work to AI agents: 1,467 live resources, each with its original link, license and safety tags.

MCP server

io.github.miller-joe/shopify-mcp

MCP server for Shopify Admin API with a ComfyUI bridge for AI product image generation.

Needs credentialsDoes not send

https://github.com/miller-joe/shopify-mcp

Needs credentials

SHOPIFY_STORE, SHOPIFY_ACCESS_TOKEN, SHOPIFY_API_VERSION, COMFYUI_URL, COMFYUI_PUBLIC_URL, COMFYUI_DEFAULT_CKPT

Install
npx -y @miller-joe/shopify-mcp

Original content

Full originalMITSource

Republished in full under its license, with attribution. The source is authoritative.

shopify-mcp

MCP server for Shopify. Full Admin GraphQL API tooling plus an AI-driven product creation bridge via ComfyUI image generation.

The pitch

Every other Shopify MCP is a plain Admin API wrapper. This one pairs with @miller-joe/comfyui-mcp so you can say things like:

> "Create a product called *'Nebula Dreamer'*. Generate a cosmic abstract image for it, description matching the vibe, tagged astrology, status draft."

Claude then runs ComfyUI, gets an image back, creates the Shopify product, and attaches the image, all in one call.

Install

# npx, no install
npx @miller-joe/shopify-mcp \
  --shopify-store your-store.myshopify.com \
  --shopify-access-token shpat_xxx

# Docker
docker run -p 9110:9110 \
  -e SHOPIFY_STORE=your-store.myshopify.com \
  -e SHOPIFY_ACCESS_TOKEN=shpat_xxx \
  -e COMFYUI_URL=http://comfyui:8188 \
  ghcr.io/miller-joe/shopify-mcp:latest

Connect an MCP client

claude mcp add --transport http shopify http://localhost:9110/mcp

Or point your MCP gateway at the Streamable HTTP endpoint.

Configuration

CLI flagEnv varDefaultNotes
--shopify-storeSHOPIFY_STORE*(required)*my-store or my-store.myshopify.com
--shopify-access-tokenSHOPIFY_ACCESS_TOKEN*(required)*Admin API token (shpat_…)
--shopify-api-versionSHOPIFY_API_VERSION2026-04GraphQL Admin API version
--hostMCP_HOST0.0.0.0Bind host (HTTP mode only)
--portMCP_PORT9110Bind port (HTTP mode only)
--stdioMCP_TRANSPORT=stdio(unset)Speak MCP over stdio instead of HTTP. Use when launched as a subprocess by a stdio-first MCP client (Claude Desktop, mcp-inspector).
--comfyui-urlCOMFYUI_URL*(optional)*Enables bridge tools when set
--comfyui-public-urlCOMFYUI_PUBLIC_URLsame as --comfyui-urlExternal URL used for image references passed to Shopify
(no flag)COMFYUI_DEFAULT_CKPTsd_xl_base_1.0.safetensorsDefault checkpoint for bridge tools

Transports

The server speaks streamable HTTP by default (great for Claude Code, MetaMCP, raw fetch). Pass --stdio (or set MCP_TRANSPORT=stdio) to switch into stdio mode, which is what stdio-first clients like Claude Desktop and the MCP Inspector expect:

// claude_desktop_config.json
{
  "mcpServers": {
    "shopify": {
      "command": "npx",
      "args": ["-y", "@miller-joe/shopify-mcp", "--stdio"],
      "env": {
        "SHOPIFY_STORE": "my-store.myshopify.com",
        "SHOPIFY_ACCESS_TOKEN": "shpat_…"
      }
    }
  }
}

Getting a Shopify access token

Easy path, existing dev store: Shopify Admin → Apps → *Develop apps* → Create custom app → enable relevant Admin API scopes (write_products, read_orders, write_inventory, read_customers) → install → copy the admin API access token (starts with shpat_).

For new apps (post-Jan 2026): legacy custom-app tokens are deprecated for freshly-created apps. Use the Dev Dashboard token-exchange flow once to obtain a working token, then supply it here. Multi-tenant OAuth is on the roadmap.

Tools

Core Admin

ToolDescription
list_productsPaginated product search with Shopify query syntax
get_productFetch one product with variants, images, media
create_productCreate a product (default DRAFT); optionally attach images
update_productUpdate title, description, tags, status, etc.
upload_product_imageAttach a public image URL to an existing product
list_ordersList orders, newest first, with query filters
get_orderFetch one order with line items
set_inventory_quantitySet absolute on-hand inventory at a location
list_locationsList store locations (for inventory ops)
list_customersList customers with query filters

Metafields

ToolDescription
set_metafieldUpsert a metafield on any HasMetafields resource (product, variant, collection, customer, order, shop, etc.)
list_metafieldsList metafields for a resource, optionally filtered by namespace
delete_metafieldDelete a metafield by (ownerId, namespace, key)

Draft orders

ToolDescription
list_draft_ordersList draft orders with Shopify query filters
get_draft_orderFetch one draft order with its line items
create_draft_orderCreate a draft order. Line items can be variant refs or custom (title + price).
update_draft_orderUpdate customer, line items, tags, note, email
complete_draft_orderConvert a draft order to a real order. paymentPending skips capture.
delete_draft_orderDelete a non-completed draft order

Webhooks

ToolDescription
list_webhooksList webhook subscriptions; filter by topic(s)
get_webhookFetch a single subscription
create_webhookSubscribe an HTTPS callback URL to a topic (e.g. ORDERS_CREATE)
update_webhookChange callback URL, format, or field/metafield filters
delete_webhookDelete a subscription

Metaobjects

ToolDescription
list_metaobject_definitionsDiscover metaobject types (schemas) on the store, including field definitions
list_metaobjectsList metaobjects of a given type
get_metaobjectFetch one metaobject with all its fields
create_metaobjectCreate a metaobject (type must already exist as a definition). Supports ACTIVE/DRAFT status.
update_metaobjectUpsert fields, change handle, toggle publishable status
delete_metaobjectDelete a metaobject

Fulfillment

ToolDescription
list_fulfillment_ordersList an order's fulfillment orders (one per shipping location), with remaining quantities per line item
get_fulfillment_orderFetch a single fulfillment order
get_fulfillmentFetch a single fulfillment (shipment record) with tracking info
create_fulfillmentMark fulfillment orders (or specific quantities) as fulfilled. Optionally attach tracking and notify the customer.
update_fulfillment_trackingUpdate carrier/number/url on an existing fulfillment
cancel_fulfillmentCancel a fulfillment by ID

Partial fulfillment is supported. Pass specific fulfillmentOrderLineItems with quantity per line, or omit the array to fulfill everything on the fulfillment order.

Variants and product options

ToolDescription
list_variantsList all variants of a product with selected options, price, SKU, inventory
create_variantsBulk-create variants (up to 100) with option values, price, SKU, compareAtPrice, initial inventory
update_variantsBulk-update variant price, compareAtPrice, SKU, barcode, taxable, inventoryPolicy, option values
delete_variantsBulk-delete variants from a product
reorder_variantsSet 1-indexed positions for variants
add_product_optionsAdd options (Size / Color / etc.) with their possible values. Up to 3 options per product.

For an entirely new product, creating the first real variant requires strategy="REMOVE_STANDALONE_VARIANT" to replace the auto-generated "Default Title" variant.

Collections and tagging

ToolDescription
list_collectionsList collections with query filters
get_collectionFetch one collection with its products
create_collectionCreate a manual collection, optionally seeded with products
update_collectionUpdate title, description, or handle
delete_collectionDelete a collection
add_products_to_collectionAdd products to a manual collection (async job on Shopify's side)
remove_products_from_collectionRemove products from a manual collection
add_tagsAdd tags to any taggable resource (Product, Order, Customer, DraftOrder, Collection)
remove_tagsRemove tags from a taggable resource

Analytics (ShopifyQL)

ToolDescription
run_shopifyql_queryRun a ShopifyQL query and render the result as an ASCII table. Pass raw=true for the raw JSON payload.

Examples:

  • FROM sales SHOW total_sales BY day SINCE -30d TIMESERIES
  • FROM products SHOW product_title, quantity_sold BY product_id SINCE -7d ORDER BY quantity_sold DESC LIMIT 10

ComfyUI bridge (when COMFYUI_URL is configured)

ToolDescription
generate_and_create_productGenerate an image and create a product with it, in one call. Title and description derive from the prompt if not given.
generate_product_imageGenerate an image and attach it to an existing product.
refine_product_imageRun img2img on a product's featured image (or an explicit URL) and attach the refined result. Tune denoise (0–1) for how far the result drifts from the source.
bulk_regenerate_imagesFor all products matching a query, run the generator with a templated prompt and attach fresh images.

Template placeholders for bulk_regenerate_images: {title}, {handle}.

Example: the whole pitch in one call

Claude, use generate_and_create_product:
  prompt: "minimalist sunset mountain silhouette, warm gradient, vector style"
  title: "Mountain Sunset Poster"
  status: DRAFT
  tags: ["posters", "nature", "minimalist"]

ComfyUI generates the image, Shopify creates the product with the image attached, and you get the product ID and image URL back. One prompt, one call, real listing.

Architecture

┌────────────────┐     ┌──────────────────┐     ┌─────────────────┐
│  MCP client    │────▶│  shopify-mcp     │────▶│  Shopify Admin  │
│  (Claude etc.) │◀────│  (this server)   │◀────│  GraphQL API    │
└────────────────┘     └────────┬─────────┘     └─────────────────┘
                                │
                                │ (bridge tools only)
                                ▼
                       ┌──────────────────┐
                       │  ComfyUI         │
                       │  (txt2img)       │
                       └──────────────────┘

Bridge tools call ComfyUI directly over HTTP, get an image URL, and pass it to Shopify's productCreateMedia mutation. Shopify fetches and hosts the image on its CDN.

Development

git clone https://github.com/miller-joe/shopify-mcp
cd shopify-mcp
npm install
npm run dev   # hot reload via tsx watch
npm run build
npm run typecheck
npm test

Requires Node 20+.

Roadmap

Shipped:

  • Core products CRUD plus image attach
  • Orders read, Inventory set + locations, Customers read
  • ComfyUI bridge: generate_and_create_product, generate_product_image, bulk_regenerate_images
  • Metafields: set_metafield, list_metafields, delete_metafield
  • Draft orders: create / update / complete / delete / list / get
  • Collections and tagging: CRUD, product add/remove, add_tags / remove_tags
  • Variants and product options: bulk create / update / delete / reorder plus add_product_options
  • Fulfillment: list/get fulfillment orders, create fulfillment (partial supported), update tracking, cancel
  • Webhooks: list / get / create / update / delete
  • Metaobjects: definitions list plus metaobject CRUD
  • ShopifyQL analytics: run_shopifyql_query with ASCII-table rendering
  • Image refinement bridge: refine_product_image (ComfyUI img2img on product images)

Planned:

  • OAuth token-exchange flow for new-app auth.

License

MIT © Joe Miller

Support

If this saves you time, consider supporting development:

DescriptionMCP server for Shopify Admin API with a ComfyUI bridge for AI product image generation.
Version0.1.2
Packagesnpm · @miller-joe/shopify-mcp@0.1.2 · streamable-http
Environment variablesSHOPIFY_STORE required
Store domain, e.g. 'my-store' or 'my-store.myshopify.com'
SHOPIFY_ACCESS_TOKEN required secret
Admin API access token (shpat_...). See README for how to create a custom app on a dev store.
SHOPIFY_API_VERSION
Shopify GraphQL Admin API version (default: 2026-04)
COMFYUI_URL
ComfyUI HTTP URL for bridge tools. When unset, generate_* bridge tools are disabled.
COMFYUI_PUBLIC_URL
Externally-reachable ComfyUI URL (defaults to COMFYUI_URL)
COMFYUI_DEFAULT_CKPT
Default ComfyUI checkpoint filename for bridge tools (default: sd_xl_base_1.0.safetensors)

Related

TutorialCustomer service

2_1_support_ticket_agent

A structured output agent demonstrating Pydantic schema-based responses for customer support ticket creation.

Sourcegithub.com/Shubhamsaboo/awesome-llm-apps/tree/main/ai_agent_framework_crash_course/openai_sdk_crash_course/2_structured_output_agent/2_1_support_ticket_agent

No credentialsDoes not sendApache-2.0
TutorialCustomer service

3_1_customer_support_ticket_agent

A tutorial demonstrating how to implement a structured customer support ticketing system using Google's ADK (Agent Development Kit) framework. This example shows how to create type-safe, structured support tickets with priority levels, categories, and resolution estimates using Pydantic schemas and

Sourcegithub.com/Shubhamsaboo/awesome-llm-apps/tree/main/ai_agent_framework_crash_course/google_adk_crash_course/3_structured_output_agent/3_1_customer_support_ticket_agent

No credentialsDoes not sendApache-2.0
CaseCustomer service

ai_customer_support_agent

This Streamlit app implements an AI-powered customer support agent for synthetic data generated using GPT-4o. The agent uses OpenAI's GPT-4o model and maintains a memory of past interactions using the Mem0 library with Qdrant as the vector store.

Sourcegithub.com/Shubhamsaboo/awesome-llm-apps/tree/main/advanced_ai_agents/single_agent_apps/ai_customer_support_agent

No credentialsDoes not sendApache-2.0
TutorialCustomer service

Using tool_choice=required for customer service (OpenAI cookbook)

Using Tool Required for Customer Service The ChatCompletion endpoint now includes the ability to specify whether a tool must be called every time, by adding tool_choice='required' as a parameter. This adds an element of determinism to how you build your wrapping application, as you can count on a to

Sourcegithub.com/openai/openai-cookbook/blob/main/examples/Using_tool_required_for_customer_service.ipynb

No credentialsDoes not sendMIT
TutorialCustomer service

Customer service agent with tool use (Claude cookbook)

Creating a Customer Service Agent with Client-Side Tools In this recipe, we'll demonstrate how to create a customer service chatbot using Claude 3 plus client-side tools. The chatbot will be able to look up customer information, retrieve order details, and cancel orders on behalf of the customer. We

Sourcegithub.com/anthropics/claude-cookbooks/blob/main/tool_use/customer_service_agent.ipynb

No credentialsDoes not sendMIT
Awesome listCustomer service

500 AI Agents Projects (includes customer service use cases)

Curated collection of AI agent use cases across industries (incl. Customer Service, E-commerce) with links to open-source implementations, e.g. a LangGraph customer-support agent tutorial.

Sourcegithub.com/ashishpatel26/500-AI-Agents-Projects

No credentialsDoes not sendMIT