1Auth
Authenticate principal
Gateway
One managed exchange layer between AI and your MSP stack.
MSPlex is not just a collection of connectors. The gateway is the product boundary that handles routing, connector auth state, and tenancy proof so AI can operate against real MSP systems without each workflow building its own infrastructure.
Architecture
The gateway is the control plane, not just a pass-through.
$ msplex connect halopsa --tenant acme
-> Gateway session ready
$ msplex status
-> operational
-> 5 connectors
-> https://msplex.ai/mcp/v1/acme-corp
Enforcement chain
Before a tool runs, the gateway clears the request.
A tool call does not go straight to a connector. The gateway validates identity, entitlements, policy, and execution state first, which is why the exchange is useful as an operational control plane instead of just a transport layer.
POST /mcp -> tools/call
Identity
2Tenant
Resolve tenant context
3Customer
Resolve customer context
Entitlements
4Snapshot
Load entitlement snapshot
5Subscription
Verify subscription status
6Connector
Verify connector entitlement
7Readiness
Verify instance readiness
Policy
8Policy
Evaluate tool policy and RBAC
9Quota
Check quota and overage
10Approval
Check approval requirement
Execution
11Dispatch
Dispatch to worker
12Audit
Write audit and metering
13Response
Return response
What the gateway owns
The exchange becomes useful when it carries the operational burden.
Auth stays in the exchange
The gateway owns token refresh, connector auth state, and the boundary between AI clients and production systems.
One route layer, many tools
Routing logic lives in one place, so AI clients do not need custom point-to-point integrations for every MSP system.
Tenant proof travels with the call
Every request carries tenant proof and stays inside a boundary that is easier to reason about than scattered MCP glue.
Evaluation rules
A better gateway review ties architecture to rollout reality.
01
Map the first workflow
Start by identifying which systems the first workflow must touch and whether the current build already covers them.
02
Review the gateway in context
Keep the gateway conversation tied to trust review, connector fit, and rollout timing instead of treating it as abstract architecture.
03
Choose a durable control plane
Use the gateway as the control boundary for production AI workflows rather than letting each agent invent its own connector logic.
Review next
Pair the gateway page with trust and connector fit.
Review next
Practical note
The gateway matters because it becomes the operational contract between AI workflows and production systems. That contract should be understandable before rollout begins.
Need the gateway explained against your current stack?
Bring the tools you want AI to reach, the trust constraints you need to preserve, and the rollout timeline you are targeting.