Fynd Model Context Protocol
Give AI tools secure access to your store context over Model Context Protocol (MCP). Set it up and start building, querying and managing commerce for your store.
Why Model Context Protocol
Model Context Protocol (MCP) is the open standard that lets AI tools connect to live data and take real actions. Fynd Storefront MCP gives AI access to one store’s catalog, cart, checkout, policies and customer journeys. Fynd Platform MCP gives authorised teams access to manage catalog, inventory, orders, offers, delivery and support using AI.
Choose your MCP
Power AI shopping experiences with Storefront MCP, or manage commerce at scale with Platform MCP.
Help customers discover products, manage a cart, complete checkout, understand store policies, track orders, start returns, and access their account after sign-in.
Ground natural-language requests in the live product catalog.
Connect authorised AI experiences to 400+ GraphQL-powered commerce capabilities across products, inventory, customers, orders, delivery, content, and support.
Prepare, validate and publish complete product assortments.
Platform MCP
Fynd groups 400+ commerce capabilities into the areas AI works in: catalog, inventory, orders, offers, customers, stores, delivery and support.
Products, variants, categories, attributes, media and channel readiness.
Understand the complete state of an assortment before deciding what should change.
Availability by SKU, store, location, selling channel and fulfilment mode.
See where stock is, where demand is building and which promises are exposed before taking action.
Order, payment, fulfilment, shipment and exception state in one operating context.
Trace an order across its lifecycle and identify the exact point preventing successful fulfilment.
Eligibility, rules, conflicts, usage and commercial constraints.
Evaluate a promotion against customer eligibility, active offers, margin rules and available stock.
Profiles, segments, behaviour, account history and service context.
Build a relevant customer view without detaching behaviour from orders, benefits and service history.
Location configuration, serviceability, availability and omnichannel operations.
Connect what each store can sell and fulfil with the demand and inventory around it.
Promise, fulfilment, shipment, carrier and exception status.
Detect delivery risk early and relate it back to inventory, fulfilment and customer impact.
Customer, order, payment, return and policy context for complete resolution.
Move from a customer question to the full commercial and operational context required to resolve it.
How it works
Use Storefront MCP for one store, or Platform MCP for company-wide commerce operations.
Use the setup prompt or command to add the Fynd MCP endpoint to a supported AI tool.
Choose the store, company and permissions the AI tool is allowed to use.
Search products, build a cart, check an order or work with an authorised platform operation.
Control by design
MCP is an access layer, not a permission bypass. Every action inherits your Fynd security model. Exact scopes, excluded fields and approval-required operations are documented for each enabled environment.
Every request stays inside the same company, application and role boundaries already configured in Fynd.
Read and analysis can run immediately. Governed writes surface the proposed change, affected records and expected impact first.
Orders, addresses, returns and account information are available only within the authenticated shopper session.
Sensitive or unnecessary fields can be excluded from the model-facing response while the underlying commerce operation remains deterministic.
Set up Fynd MCP
Choose your AI tool and copy the setup to connect.
Paste this into an AI tool that can add MCP connections from chat.
It is a Model Context Protocol connection that makes Fynd commerce capabilities understandable and usable by MCP-compatible AI clients.
No. Every Fynd store exposes a built-in public endpoint at its own domain. Customer-specific information and actions still require the customer to sign in.
It can help a customer discover products and build a cart, then continue into the store's checkout experience. Checkout behaviour follows the storefront's existing flow.
Authentication is required before personal order details, tracking, return actions, or account information are made available.
It is for teams, developers, and authorised applications building AI experiences that manage commerce operations across Fynd Platform.
No. Storefront MCP exposes public shopping capabilities and protects customer-specific data with sign-in. Platform MCP uses authorised platform access appropriate to the connected context.
