About shopbuddy
Shop Buddy is a hosted service and MCP server for coding agents that make Shopify theme changes. It uses unpublished staging themes, Theme Check, before-and-after screenshots, owner approval, and rollback rather than editing the live theme directly.
Key features
- 01Edit an unpublished staging theme
- 02Run Theme Check validation
- 03Capture mobile and desktop screenshots
- 04Summarize visual changes by section
- 05Request owner approval to publish
- 06Rollback changes with one tap
Use cases
- 01Review an agent's theme edits before publishing
- 02Validate theme files with Theme Check
- 03Compare mobile and desktop previews
- 04Undo a published change
MCP support
The hosted MCP server lets compatible clients connect Shop Buddy to coding agents. Its tools manage staging edits, validation, previews, approval, publishing requests, and rollback.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://getshopbuddy.vercel.app/api/mcpclaude mcp add --transport http tmill313-shopbuddy https://getshopbuddy.vercel.app/api/mcp{"mcpServers":{"tmill313-shopbuddy":{"type":"http","url":"https://getshopbuddy.vercel.app/api/mcp"}}}Open in Cursor{"servers":{"tmill313-shopbuddy":{"type":"http","url":"https://getshopbuddy.vercel.app/api/mcp"}}}Open in VS Code{"mcpServers":{"tmill313-shopbuddy":{"type":"http","url":"https://getshopbuddy.vercel.app/api/mcp"}}}{"mcpServers":{"tmill313-shopbuddy":{"serverUrl":"https://getshopbuddy.vercel.app/api/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
Tools
Details
From the project READMEShop Buddy
Let your AI coding agent change your Shopify theme safely. Every edit lands on an unpublished staging theme. Shop Buddy validates it with Theme Check, screenshots before/after on mobile and desktop with a plain-language summary of what changed, and you approve with one tap. Rollback is one tap. The live theme is never edited directly.
Site and one-click install: https://getshopbuddy.vercel.app/install
# Claude Code (hosted MCP, signs in through your browser)
claude mcp add --transport http shopbuddy https://getshopbuddy.vercel.app/api/mcp
# Claude Code plugin (adds the skill too)
claude plugin marketplace add tmill313/shopbuddy-plugin
claude plugin install shopbuddy@shopbuddy-marketplace
# Codex
codex mcp add shopbuddy --url https://getshopbuddy.vercel.app/api/mcp
# Local CLI (no account needed for local mode)
npx shopbuddy
Claude (web/desktop) and ChatGPT: add a custom connector with https://getshopbuddy.vercel.app/api/mcp. The agent opens a browser tab; sign in or create an account there and click Allow. No keys.
How it works
shopbuddy_start_changeduplicates the live theme into a staging theme and pulls it into a server-side working copy.- The agent reads, searches and writes files, runs Theme Check with
shopbuddy_validate, thenshopbuddy_push_previewuploads to staging and captures before/after screenshots, each visual change attributed to its section. shopbuddy_submitandshopbuddy_request_publishhand the store owner a one-tap approval link. Owners can opt a store into trusted publishing so low-risk changes publish on request, always logged and one tap fromshopbuddy_rollback.
Agent-readable summary: https://getshopbuddy.vercel.app/llms.txt
What is in this repo
This is the public companion to the hosted service: the Claude Code plugin and marketplace manifests, the shopbuddy skill, the MCP client config, and the MCP Registry manifest (io.github.tmill313/shopbuddy). The platform itself is a hosted service; the CLI ships compiled on npm as shopbuddy.
Privacy and terms
https://getshopbuddy.vercel.app/privacy · https://getshopbuddy.vercel.app/terms
Shop Buddy is an independent service, not affiliated with or endorsed by Shopify.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.