Redo MCP server for your AI

No. The connect form asks for two values, Store ID and API Secret, and both come from the same screen in the Redo dashboard under Settings → Developer. There is no OAuth app, client secret or redirect URL to register. The secret is scoped to a single store, so the connection is locked to that store.

44 operations across 10 API modules, split 23 read and 21 write. 25 load into your AI client by default and the rest stay searchable, so Redo doesn't crowd out the other servers in your tool list. Returns is the largest module with 12 operations, followed by Webhooks, Inventory and Orders.

Yes. process_return triggers the real outcome for the products you pick, and Redo marks that operation BETA. update_return_status only moves a return's status; it doesn't move money. Read-only mode blocks the money-moving call outright, and approval-required-for-writes holds it for a person to confirm first.

Yes. create_order ingests an order into Redo so returns and exchanges can be made against it, which is the usual path for stores Redo doesn't sync natively. import_returns brings across returns that were created outside Redo. Storefront browsing and checkout events go in through create_storefront_event, which expects the Shopify web-pixel event schema.

Both. list_inventory_items covers in-stock items with on-hand quantity above zero, and list_inventory_levels gives the aggregate quantity per product or variant at each location. On the receiving side, list_inbound_shipments covers returns and transfers heading to a warehouse, and get_inbound_shipment adds that shipment's purchase orders, transfer orders, files and any receiving discrepancies.

get_customer looks a customer up by email, and email is a required query parameter, so you need the address on hand. delete_return_tag takes the tag name in the path, case-sensitive and URL-encoded, so VIP and vip are not the same tag. create_webhook upserts by externalId, which means reusing an externalId updates the existing subscription instead of creating a duplicate.

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 44 tool definitions into every conversation burns context before the work starts. PopMCP loads a curated 25-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 44 here. A trigger is something PopMCP watches for you: 12 Redo 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 Redo.