The Scope3 Storefront API provides capabilities for publishers and technology partners to set up and manage storefronts, configure inventory sources, register agents, and handle billing through both REST and MCP (Model Context Protocol) interfaces.
Manage storefronts for your publisher integration. Each customer has exactly one storefront, which is automatically created in PENDING when the seller signs up — POST /api/v2/storefront is idempotent and returns the existing storefront if one is already provisioned. There is no delete endpoint; storefronts are activated or disabled via PUT.
Method
Endpoint
Description
GET
/api/v2/storefront
Get your storefront
PUT
/api/v2/storefront
Update storefront (also used to activate / disable)
POST
/api/v2/storefront
Idempotent: returns the existing storefront, creates one if missing
Manage inventory sources within a storefront. To register an agent, create an inventory source with executionType: "agent" and provide the agent’s endpoint URL and protocol.
Method
Endpoint
Description
GET
/api/v2/storefront/inventory-sources
List inventory sources
POST
/api/v2/storefront/inventory-sources
Create inventory source
GET
/api/v2/storefront/inventory-sources/:sourceId
Get inventory source
PUT
/api/v2/storefront/inventory-sources/:sourceId
Update inventory source
DELETE
/api/v2/storefront/inventory-sources/:sourceId
Delete inventory source
Example — Register an Agent as an Inventory Source:
The Storefront MCP server uses a generic tool pattern. AI agents interact with the full REST API surface through these tools:
Tool
Description
ask_about_capability
Query the API documentation to look up correct endpoint paths, field names, and request schemas. Call this before api_call when unsure of the exact interface.
api_call
Authenticated HTTP proxy to any storefront REST endpoint. Supports all methods (GET, POST, PUT, PATCH, DELETE) with full request/response passthrough.
health
Health check to verify the API is reachable.
Required workflow for AI agents: Always call ask_about_capability first to retrieve accurate field names and schemas before calling api_call. Field names are camelCase (e.g. agentId, not agent_id).