Okendo MCP server for your AI

No. The connect form asks for one thing, your Okendo User ID. The Storefront REST API this integration reads is public, so there is no key to generate or rotate. The User ID is what pins the connection to a single store, so a second store means a second connection with its own User ID.

Yes. Aggregates, review lists, customer media and Q&A all come in product and collection versions, where a collection is what Okendo treats as a review group. An agent can pull the star distribution for one SKU with get_product_review_aggregate, or roll up an entire Shopify collection with get_collection_review_aggregate and list_collection_reviews.

No. Reviews, rating aggregates, customer media, Q&A, AI summaries and loyalty configuration are all read paths, so nothing here publishes, hides, or rewrites content. The only calls that write are create_referral_sender, which registers a shopper as a referral sender and returns their referralSenderId and share link, and loyalty_login. Moderation and review responses stay in the Okendo admin.

No. Everything the agent gets back is published content, the same reviews, media and Q&A a shopper sees on your storefront. This is the detail that catches people out: a count from list_reviews or list_product_questions describes what is live, not what is sitting unpublished. Judge the numbers against your storefront rather than your moderation queue.

loyalty_login exchanges a logged-in shopper's Shopify session token for a short-lived Okendo loyalty JWT plus that shopper's loyalty details. It needs a live session token, so an agent cannot use it to browse shopper accounts at will. The other loyalty calls return program configuration instead: earning rules, redemption rules, VIP tiers and their requirements, the published FAQ, and the legal and enrollment copy.

Yes, that is what the triggers are for. Threshold watches fire when the store-wide average rating drops below a level you set, when a single product average drops, or when total published reviews cross a milestone. Change detection covers a negative review going live, a new review on a specific product or collection, a shopper adding a photo or video, a new question, and edits to loyalty earning rules, rewards or VIP tiers.

No. PopMCP hosts it and gives you a Streamable HTTP MCP URL. There is nothing to deploy, no callback URL to register, and no infrastructure to keep running or patch.

Any client that supports a remote Streamable HTTP MCP server — Claude Desktop, Claude Web, ChatGPT custom connectors, Codex and the Codex CLI, OpenCode and Z.ai's ZCode, as well as editors such as Cursor, Windsurf and Zed. Your client signs in to PopMCP with OAuth; the provider credential never leaves the server.

Loading 20 tool definitions into every conversation burns context before the work starts. PopMCP loads a curated 20-tool set covering the everyday workflows and keeps the rest searchable, so your AI pulls a specialist tool only when a task needs it. You can switch a connection to the full catalog at any time.

A tool is something your AI calls — there are 20 here. A trigger is something PopMCP watches for you: 13 Okendo events that can start work on the PopMCP side. Triggers never enter your AI client's context, and none of them ask you to paste a callback URL into Okendo.