Built for Shopify Say "optimize my worst products" and watch it happen.
StoreSEO exposes your store to AI assistants two ways: a remote MCP server at mcp.storeseo.com for Claude and other clients that support custom connectors, and the AI Toolkit with API tokens for Claude Code, Cursor and VS Code. Either way your SEO work stops being a screen you visit and becomes a sentence you type.
StoreSEO is the AI SEO Agent for Shopify merchants — a Shopify app that optimizes your store for Google and gets it cited by ChatGPT, Gemini, Claude and Perplexity.
4.9/5 from 650+ Shopify reviews
What you get
- OAuth 2.0
- secured remote MCP connection
- Claude
- Cursor and VS Code via AI Toolkit
- Sidekick
- reads StoreSEO inside Shopify admin
The problem
The work is mechanical but not automated
Finding low-scoring products, rewriting meta fields, adding alt text and re-checking is a loop a person should not be running by hand hundreds of times.
Context lives in another tab
You are already in an AI assistant planning content or analyzing data. Switching to an admin panel to act on that analysis breaks the thread every single time.
Nobody audits on a schedule
SEO decays quietly. New products ship unoptimized, descriptions drift, scores fall — and nothing surfaces it until traffic does, months later.
How the agent surface works
Six things that exist specifically so an AI assistant can run your SEO rather than describe it.
StoreSEO MCP server
A remote MCP server at mcp.storeseo.com/mcp. Add it as a custom connector in Claude, authorize the store, and every SEO action becomes available in conversation, secured with OAuth 2.0.
Connect the MCP serverAI Toolkit and API tokens
Generate a named token in Settings and connect Claude Code, Cursor or VS Code from the terminal. Same capabilities, aimed at people who live in a shell.
Set up the AI ToolkitShopify Sidekick integration
Shopify's own assistant can read your StoreSEO issues, open the exact fix page in admin and run the AI content generator without leaving the conversation.
Use with Sidekickagents.md editor
The instructions AI shopping agents read about your store, editable without touching a theme file, so agent traffic gets your guidance rather than a generic default.
Edit agents.mdLLMs.txt generator
Publishes the curated map answer engines read, versioned with an active-file badge so rolling back to an earlier file takes one click.
Generate LLMs.txtAuto AI optimization
The standing rule underneath the conversational layer: new products get optimized on arrival, so the agent has less to clean up next time you ask.
Automate itWays people use it
Concrete situations, tagged by who hits them. Each one names the trigger and what to set up.
Optimizing from the terminal
You live in a shell and do not want another admin UI in the loop.
When
Any time you would rather type than click.
Set up
Generate an AI Toolkit token and add the MCP server to Claude Code.
SEO as part of a deploy routine
New products ship and nobody remembers the SEO pass.
When
On every catalog change.
Set up
Enable auto optimization and use the agent for periodic verification runs.
You do not know what to ask for
The app has plenty of features and no obvious starting point.
When
First week with the app.
Set up
Connect the MCP server and ask the assistant to audit the store and propose a plan.
Answering Sidekick questions
You are already talking to Shopify's assistant about a product.
When
While working inside Shopify admin.
Set up
Enable the Sidekick integration so it can read StoreSEO issues directly.
Running many stores from one place
Ten client stores, ten admin logins, ten separate audits.
When
Reporting week.
Set up
Connect each store over MCP and run the same audit prompt against all of them.
Agentic commerce readiness review
Clients ask whether AI shopping agents can transact with their store.
When
As part of a quarterly review.
Set up
Check the agentic discovery score, publish LLMs.txt and edit agents.md.
The prompts
Connect your store via the StoreSEO MCP server or the AI Toolkit, then paste any of these into Claude. They are written for this kind of store, not as a generic template.
Using StoreSEO, take ownership of my store's SEO. Audit every product, collection and page, fix everything fixable — meta titles, descriptions, URL handles, image alt text, FAQ schema and AI snippets — without changing my brand voice. Then every Monday, re-audit and fix any regression, and post me a summary of what changed and what still needs a human.
Using StoreSEO, make my store readable to AI answer engines. Generate AI snippets for my top products, publish an up-to-date LLMs.txt, rewrite agents.md so shopping agents describe my brand accurately, add FAQ schema where it is missing, and check my agentic discovery score. Tell me what improved and what is still weak.
Using StoreSEO, list my 50 lowest-scoring products with their current SEO score, then optimize all of them: focus keyword, meta title, meta description, URL handle, image alt text and FAQ schema. Re-score each one afterwards and show me a before and after table sorted by biggest improvement.
Run it from Claude
What actually happens when you ask. An illustrative trace of the tool calls the agent makes on your behalf.
> SEO optimize all my least scoring products.
- · storeseo.list_products — sorted by SEO score ascending
- · storeseo.get_seo_score — 214 products below the threshold
- · storeseo.optimize_product — meta title, description and handle per product
- · storeseo.generate_alt_text — images described
- · storeseo.add_faq_schema — questions added where the description supports them
- · storeseo.get_seo_score — re-scored to confirm the change
A summary table of every product with its score before, its score after, and exactly what was fixed.
In your industry
6 businesses that run this kind of store, each with its own story and its own prompt.
Questions
How do I connect StoreSEO to Claude?
In Claude, open Settings, go to Connectors, choose Add Custom Connector, and enter the StoreSEO MCP server URL https://mcp.storeseo.com/mcp. You will then be asked to connect and choose the store you want to optimize, and to set what the assistant is allowed to do. After that you can run SEO tasks by typing plain-English prompts.
What is the difference between the MCP server and the AI Toolkit?
The MCP server is a remote connector you add inside an AI assistant such as Claude, authorized over OAuth. The AI Toolkit generates API tokens in StoreSEO Settings for connecting local coding assistants like Claude Code, Cursor and VS Code from your terminal. Both reach the same SEO capabilities; pick whichever matches where you work.
Can the AI change things in my store without asking?
You control what the assistant is allowed to do when you authorize the connection, and you can review actions as they happen. Start by asking for an audit rather than a bulk rewrite if you want to see the plan before anything changes.
Do I need to know anything technical to use this?
For the MCP connector, no. It is a settings screen and a URL, then you type requests in ordinary English. The AI Toolkit route assumes basic comfort with a terminal because it is aimed at coding assistants.