Stormly's MCP connector gives ecommerce teams a direct bridge between their analytics data and the AI tools they already use. Connect Claude, Claude Code, ChatGPT Developer Mode, Codex, Cursor, or another MCP-compatible client, then ask plain-language questions about your store performance.
"Which products drove the highest repeat purchase revenue last month, and which had high cart abandonment?"
Built for Ecommerce Questions, Not Generic BI
Ecommerce analytics is full of relationships that generic dashboards miss: products, variants, brands, collections, carts, orders, returns, campaigns, and repeat buyers. Stormly keeps that context in the analytics layer, then exposes it through MCP so your AI client can ask the right questions.
Product and SKU analytics
Ask about SKU revenue, variant performance, category lift, cross-sell behavior, bundles, and new-arrival performance.
Cart and order analytics
Analyze cart abandonment, recoverable revenue, checkout starts, order conversion, AOV, and revenue contribution.
Funnels and retention
Break down product views, add-to-cart behavior, purchase funnels, repeat buyers, cohorts, and returning customer segments.
AI-ready answers
Stormly submits the question, runs the analysis, then returns a Markdown answer based on actual project data.
How the Stormly MCP Connector Works
- Connect your store data to Stormly through Shopify, Adobe Commerce, Segment, Google Tag Manager, React, Wix, CSV, or an API.
- Create an MCP setup token or connect through OAuth from your Stormly project settings.
- Add the project-scoped MCP endpoint to Claude, Claude Code, ChatGPT Developer Mode, Codex, Cursor, or another MCP client.
- Ask one complete ecommerce analytics question, including date ranges, filters, dimensions, and comparison periods.
- Stormly chooses a matching report or SQL path, runs it against the project data, and returns the interpreted answer.
Questions You Can Ask
- Which SKUs grew revenue fastest last week compared with the previous week?
- Which products have the highest add-to-cart rate but weak checkout completion?
- How did paid search shoppers behave differently from email shoppers in the purchase funnel?
- Which brands created the most repeat-buyer revenue in the last 90 days?
- What changed in AOV, conversion rate, and units sold after a campaign launch?
- Which product categories have the most recoverable cart abandonment revenue?
Why Ecommerce Teams Use MCP With Stormly
MCP is useful only when the server behind it understands the business domain. Stormly already models ecommerce behavior, product performance, funnels, orders, revenue, and customer retention. The MCP connector makes that model available inside the AI client where teams ask follow-up questions, compare periods, and turn analysis into actions.
Frequently Asked Questions
Is this only for Claude?
It works with MCP clients that can connect to remote HTTP MCP servers with OAuth or bearer-token headers. The current setup copy covers Claude, Claude Code, ChatGPT Developer Mode, Codex, Cursor, and other compatible clients.
Can it answer questions from actual Stormly reports?
Yes. Stormly can route an MCP question to matching ecommerce reports or to custom SQL, depending on which path best fits the request.
Is access project scoped?
Yes. Each MCP endpoint and credential belongs to one Stormly project, so the client only works with data from that project.
Ready to connect an AI client? Follow the MCP setup guide.