The CRS team publishes an MCP (Model Context Protocol) server that exposes the Customer-portal API as a set of AI-callable tools. Plug it into an MCP-aware AI client (Claude Code, Claude Desktop, Cursor, and others) and you can ask in plain English to look up users, inspect hosts, run queue operations, query Offers IQ logs, and more — without writing curl commands or building admin scripts. It's a productivity layer for your operations and support teams; it does not change anything the API itself can't already do.
The package is published on npm as @crscreditapi/b2c-mcp-customer. It runs locally via npx, authenticates with your Customer Portal credentials, and translates AI prompts into Customer-API calls.
- Triage and support — investigate a single user's identity-verification or enrollment state by name without crafting requests.
- Bulk inspection — list users, hosts, rule sets, or audit logs filtered by criteria you describe.
- Operational tasks — pause/resume/close batches of users, rotate or inspect host configuration, monitor queues.
- Exploration during integration — verify what your account contains while you're building against the API.
This MCP is customer-portal scoped — every call runs under the authenticated customer's permissions, so you can only see and act on data that belongs to your customer account.