Riggr connects your Shopify store to any AI tool that speaks MCP, Claude Code, Cursor, Codex, Claude Desktop, Claude.ai, and ChatGPT. Instead of copy-pasting data between your Shopify Admin and an AI chat window, your AI reads your live store data directly and helps you build custom apps for it.
You attach a Spark on a Riggr project, copy Talk to your AI, and paste it into your coding tool. Your AI reads your Shopify data through Riggr, scaffolds a working web app on your computer, and you iterate from there. That's the whole loop.
Riggr is a fintech-grade data layer between your Shopify store and any AI tool that speaks the Model Context Protocol (MCP). Under Shopify Protected Customer Data Level 1 it exposes read-only Shopify ops data across orders, fulfillments, products, returns, locations, shipping, and sales-channel publications — through secure read-only functions your AI can call in real time, without customer name, email, phone, or street address. Extra tools let the AI manage Riggr projects on your behalf. Customer contact fields and other contact-heavy resources need Protected Customer Data Level 2; additional read scopes (inventory levels, analytics, payments data) unlock as Shopify approves them.
The point isn't the connection itself. The point is what your AI does with it: build actual custom software for your Shopify business. Not chat responses. Not summaries. Real apps you can use every day.
I understand what Riggr is
Acknowledge below so we know this landed.
An open standard, backed by Anthropic and OpenAI, that lets AI tools reach live data sources through a common interface. Riggr is a Shopify MCP server.
A project holds up to eight Sparks, one API key, and one Talk to your AI path. Related desks share one local app (Compose + Add-ons) instead of a tab per idea. Details: Sparks & projects.
Spark is Riggr's word for a prompt. Riggr ships hand-crafted Sparks (~30KB each) that tell your AI how to scaffold specific merchant tools — returns triage, fulfillment lag, Daily Ops, and more. First two Compose into one app (emerald Do this first kickoff prompt); Sparks 3+ are Add-ons after the starter is ready, merged one at a time.
Your AI loads the Spark over MCP (via Talk to your AI), reads your Shopify data, and scaffolds a working Next.js app on your computer at localhost. You iterate with the AI to make it yours. Riggr does not host the finished app.
I know MCP vs Sparks vs the AI builder
Acknowledge below so we know this landed.
Riggr is built for Shopify merchants who want more than off-the-shelf apps. Concretely, that means:
Riggr's Sparks are designed so your AI does the coding. You need to be comfortable running a terminal command or two (open a folder, run npm run dev), but you don't need to know Next.js, React, or GraphQL. Your AI knows those.
Real requirements to run Riggr:
npm install, npm run dev).Claude Code, Cursor, or Codex if you want Autopilot to scaffold the app on your machine. Claude Desktop, Claude.ai, or ChatGPT if you want chat with live store data first (ChatGPT users hand off to Codex). Never used Terminal or Node? Start with Your computer setup.
I know the requirements to run Riggr
Acknowledge below so we know this landed.
Shopify has 2.8 million live stores worldwide, with the largest verticals being Apparel & Fashion (17.2%), Home & Garden (11.8%), and Health & Beauty (9.4%). Riggr fits merchants across all of these. Some concrete profiles:
Being explicit:
Ready to set up? Head to the Setup guide for Shopify → Subscribe → Project → Connect AI → Talk to your AI. Need Terminal / Node / npm run dev first? See Your computer setup. Or jump to Claude Code or Cursor if you already know your tool. Want to see what Riggr can build first? Check out the Sparks guide or Use cases by merchant type.