Overview
Ask questions about your reviews in plain language, and get answers, summaries and marketing ideas back in ChatGPT or Claude. The Okendo MCP Server connects your public reviews data to these AI tools, so you can uncover insights and power AI-driven shopper experiences.
What you can do with the Okendo MCP Server
You can use it to:
Summarize sentiment and themes across your product reviews.
Generate authentic marketing content using real customer voices.
Identify product opportunities based on positive and negative feedback.
Power AI tools and shopper experiences that use real customer insights.
Use it for internal analysis, marketing workflows, or partner-built shopper experiences.
Before you begin
🔔 The Okendo MCP Server is available on the Growth plan and above. View pricing.
You also need:
Published reviews from your customers.
Admin access to your Okendo dashboard.
Access to an AI tool that supports MCP connections (e.g. ChatGPT or Claude).
Retrieve your MCP URL
Open the Integrations page in Okendo, by selecting Okendo Admin > Settings > Integrations
Select the pencil icon next to the Okendo section within the AI & MCP category
Toggle on Enable MCP Server to allow the integration.
Select Copy next to your MCP URL
Then press Save to confirm the changes.
Connect the MCP Server to ChatGPT
Connect to the Okendo MCP Server
Navigate to Plugins from the sidebar, press the Add button, and select Add custom MCP server:
Configure the following fields on the Add custom MCP server dialog:
Name: Okendo
MCP Server URL:
https://mcp.okendo.io/storefront/{USER_ID}Authentication: No Auth
Check I understand and want to continue and then select Create as plugin
Select Connect Okendo to finalize the connection to the Okendo Storefront MCP Server
Install the Shopify plugin
Navigate to Plugins from the sidebar, and search for the Shopify plugin
Select Install plugin and follow the prompts
Connect the MCP Server to Claude
💡 Claude Free plans only allow adding 1 custom connector.
Connect to the Okendo MCP Server
Navigate to Settings > Customize > Connectors.
Select Add > Add custom connector
Configure the following fields on the Add custom connector dialog:
Name: Okendo
Remote MCP server URL:
https://mcp.okendo.io/storefront/{USER_ID}
Select Add and then Connect to finish adding your connector
Confirm that the MCP tool has been successfully connected by selecting the plus (+) button and then Connectors under the chat bar
Connect the Shopify connector
Navigate to Settings > Connectors > Directory, and search for the Shopify connector
Select Connect to Claude and follow the prompts
Available Okendo MCP tools
reviews_aggregate
Retrieves product review aggregate data including total reviews, rating distribution, average ratings, recommendation percentage, and attribute data for Shopify products. Can be retrieved for a specific product, or for the entire store. If the product belongs to a group (collection), aggregate data for that group is respected too.
reviews_search
Lists product reviews matching specified filter criteria with support for text search, rating filters, date ranges, media-only filtering, and pagination. Each review returned includes its media, keywords, reviewer details, sentiment, and language, and, like the other tools, respects group (collection) display settings.
questions_list
Lists published and approved product questions and their associated answers for a given Shopify product, in line with what's available in the reviews widget on your product page. If the product belongs to a group (collection), questions and answers for that group will also be retrieved. Supports sorting by most recent, oldest, or most helpful, with pagination.
A note on multi-storefront stores
The MCP mirrors your reviews widget's data, storefront by storefront. If your reviews are scoped to a single store, the MCP for that store only returns that store's reviews. If your reviews are syndicated across storefronts, the MCP reflects that same syndicated set.
A note on scope
The Okendo MCP Server is public: it only ever surfaces published, storefront-visible reviews and Q&A. It doesn't expose unpublished reviews or other private data.
Use cases for Okendo MCP Server
Merchandising and product development
Make data-driven product and merchandising decisions by turning review sentiment and themes into actionable insights.
How merchants can use it:
Ask: "What are customers saying about the fit and sizing of our new collection?"
Aggregate sentiment across products to identify manufacturing or quality issues early.
Compare new vs. legacy product sentiment to guide restocking or discontinuation.
Identify recurring keywords (e.g., "too small", "scratchy", "perfect fit") to feed back to design teams.
Detect feature requests or unmet needs that can feed into R&D.
Example queries:
"Summarize key complaints about our boots line."
"What words most commonly appear in 5-star reviews for our boots?"
Marketing and campaign planning
Help translate authentic customer feedback into marketing messaging and creative assets to drive acquisition and conversion.
How merchants can use it:
Surface top themes or sentiments to shape messaging, campaigns, and UGC use.
Identify emotional drivers of purchase decisions (e.g., "perfect for travel", "eco-friendly").
Pull real quotes and aggregate sentiment for landing pages, ad copy, or emails.
Analyze which features or benefits customers emphasize most to optimize positioning.
Example queries:
"Summarize what customers love most about our 'Chelsea Rise Vintage Tan'."
"Which products have the most positive sentiment around 'sustainability'?"
"Show me reviews that talk about this being a good 'gift'."
"Put together a list of customer quotes from my 5-star reviews that I can use in my marketing copy."
Customer support and CX
Use review data to identify, reduce, and pre-empt customer pain points, improving satisfaction and lowering support overhead.
How merchants can use it:
Identify recurring issues that drive support volume ("delayed shipping", "wrong size", "poor packaging").
Understand what customers praise or complain about most, then proactively address it in FAQs or chatbots.
Track changes in sentiment over time after product changes or campaigns.
Example queries:
"What themes appear most often in 1–2 star reviews this month?"
"Which products receive the most mentions of 'customer service'?"
"Summarize top pain points customers mention about 'shipping'."
Loyalty and retention
Use reviews as a retention layer to identify advocates, at-risk customers, and engagement opportunities for your Loyalty and Memberships programs.
How merchants can use it:
Identify promoters and advocates from reviews to invite into loyalty tiers or memberships.
Analyze churn signals from negative reviews that mention "cancelled", "too expensive", or "didn't work."
Example queries:
"Which customers mention being repeat buyers?"
"What are the main reasons customers say they stopped purchasing?"
Partners and agencies
Help provide data-backed strategic recommendations to merchants by analyzing review sentiment, trends, and themes through the MCP Server. Additionally, power conversational shopping assistants or "Store Concierge" bots using Okendo's MCP Server.
How partners use it:
Analyze merchant review data at scale, surfacing insights that inform brand, CX, and marketing strategy.
Pull Okendo review sentiment data to identify customer priorities, complaints, and trends across SKUs or collections.
Connect Okendo MCP Server to allow queries inside Storefront Concierge bots or custom GPTs that help shoppers discover products and make confident buying decisions using verified customer sentiment.
Example queries:
"Summarize the top positive and negative themes in all reviews for the last 90 days."
"Compare customer sentiment before and after our June product relaunch."
"Find reviews mentioning 'gift' or 'holiday' for products in the 'Boots' category."
"Show 3 customer quotes describing the fit of the 'Chelsea Rise Vintage Tan'."
Best practices
Be specific in your prompts: Ask clear, targeted questions for more accurate AI responses.
Filter your queries: Limit your scope by product or time range to get cleaner, more relevant insights.
Review before sharing externally: AI responses may occasionally misinterpret tone or meaning. Always verify before publishing or acting on insights.
Disclaimer
❗ Outputs generated through AI models using the Okendo MCP Server are summaries and may not always reflect exact customer statements. Always review results before public use.
FAQs & Troubleshooting
Q: What can I do with it?
A: For example:
Summarizing customer sentiment across products, now with sentiment and language returned on every review, plus reviewer details and keywords.
Identifying common likes and dislikes to inform promotions or product improvements.
Quickly generating marketing copy/snippets grounded in authentic reviews, including filtering to reviews that include photo or video media.
Answering product Q&A questions, in line with what's already shown in the reviews widget on your product page.
Powering AI concierge bots with trusted review data to help shoppers.
Q: Can shoppers use this directly on my store?
A: No. The initial release is merchant and partner-facing, and needs to be accessed with a subscription to ChatGPT or Claude. It's designed to help you analyze, summarize, and activate your public review data. However, developers and agencies can integrate the Okendo MCP server into concierge bots or other shopper-facing AI experiences.
Q: Do I need technical skills to use MCP?
A: Basic setup is straightforward, but more advanced use cases (e.g., custom concierge bots, integrations into dashboards) may require developer support. You don't need any technical skills to query the data in an AI agent like ChatGPT or Claude, though.
Q: How is this different from the Shopify Storefront MCP server?
A: Shopify's Storefront MCP server connects your live store data, such as your product catalog, variants, policies, and cart actions, directly into AI models such as ChatGPT or Claude. Okendo's MCP Server complements this by connecting your authentic public review data.
Together they enable richer, more accurate insights. For example, you can ask:
"Which products under $100 get the highest ratings for durability?"
"Highlight 5 products with the best customer feedback about quality, and give me language I can reuse in ad copy."
Q: Can partners or agencies use this on behalf of merchants?
A: Yes. Agencies and technology partners can use Okendo's MCP server to build AI-powered tools and workflows that incorporate your review data, from dashboards to concierge bots.
Q: Does connecting to the Okendo MCP Server open up any data or privacy issues?
A: No. It only reads published, storefront-visible reviews and Q&A (see A note on scope), with no access to private customer data or email addresses. If you need unpublished reviews or other private data, contact your Okendo contact to share your use case.
Q: I have multiple Shopify storefronts. Can I analyze review sentiment across all my stores at once?
A: Yes! You can set up each store individually by copying the MCP URL from each store's Okendo integration dashboard and creating individual MCP connectors in ChatGPT or Claude. Once configured, you can query them together by asking Claude to "show me the top sentiments from each store broken down." Alternatively, you can query each store separately if you prefer individual analysis.
Each store's MCP returns what that store's reviews widget shows: see A note on multi-storefront stores.
Q: What kind of data can I access with this integration? Can it pull my order data or average order value?
A: The Okendo MCP integration can only access and analyze your public review data: it searches and aggregates review information exclusively. It cannot access broader store data like order information, average order value, or other Shopify store metrics.
Q: Can I use this to help draft responses to customer reviews or support tickets?
A: Yes! You can use ChatGPT or Claude to help craft empathetic, personalized responses. For example, you could prompt: "Pull all reviews from the last 14 days that mention [specific issue]" and then ask Claude to "provide an empathetic response that includes [your specific parameters like refund offers, solutions, etc.]."
While this can save time, we recommend personalizing AI-generated responses before sending them to maintain authenticity with your customers.
Q: Can I train the AI to respond in my brand's voice or specific tone?
A: While you cannot train the underlying AI model (ChatGPT or Claude), you can control the tone and style through your prompts. For example, you can specify: "Write a response to this review in a professional and empathetic tone" or "Summarize these reviews in a casual, friendly voice." The AI will adapt its language based on your instructions within each prompt.
Q: Can I connect other data sources like Google Analytics alongside my Okendo reviews?
A: Theoretically, yes! If other platforms have MCP server integrations (like Google Analytics), you could connect multiple MCP data sources and query them together in ChatGPT or Claude. This would allow you to cross-reference review data with other business metrics. However, the effectiveness will depend on how well ChatGPT or Claude can connect insights across different data sources.