Evaluate

Evaluate KEEP

This page explains what evaluating KEEP actually involves, from before you start to how you'd end it, so there's nothing left to guess. It is a procedural reference, not a sales pitch.

Is KEEP right for your MSP?

KEEP is built for managed service providers monitoring more than one client network, and for IT teams monitoring a single network of their own. It is not a helpdesk, a backup solution, or a compliance-certifying authority. (Source: KEEP architecture documentation) For the complete picture of who KEEP is designed for, and who it isn't, see Why KEEP.

Evaluation prerequisites

A Hub deployment needs at least 4 CPU cores, 8 GB of RAM, and 100 GB of disk. A Spoke deployment (for monitoring a single client site) needs at least 2 CPU cores, 4 GB of RAM, and 32 GB of storage. (Source: KEEP architecture documentation)

Outbound internet access is required to activate and register the deployment before it begins normal operation. (Source: KEEP Evaluation Program documentation)

Before installation begins, KEEP issues an Evaluation ID and a deployment token directly to you — this isn't a self-service signup. See How KEEP Works for how this fits into the broader evaluation flow. (Source: KEEP Evaluation Program documentation)

What is installed

A KEEP Hub is installed as a software deployment, either on hardware KEEP ships or on your own hardware or virtual machine. If you're monitoring more than one client site, one Spoke is installed at each additional site. (Source: KEEP architecture documentation) Nothing is installed on the individual devices KEEP monitors. See How KEEP Works for the full architecture.

What happens during evaluation

Installation activates and registers the deployment before anything else happens — see How KEEP Works for that step in detail. Once registration succeeds, the deployment's first scan of a client network is treated as unverified evidence, not an accepted baseline, until a qualified person reviews and approves it. (Source: KEEP architecture documentation) Normal monitoring begins after that review.

Several of Merlin's AI-assisted surfaces — device diagnosis, onboarding scan analysis, and incident coordination — are fully implemented but still gathering real-world validation, which is exactly what the Founding Evaluator Program exists to do. See Merlin for what it is, and Capabilities for its current status.

What data is collected

Data collection during an evaluation works the same way it does in normal operation — nothing about being in evaluation changes what stays local or what reaches KEEP. (Source: KEEP architecture documentation) For the complete account of what stays on your own infrastructure and what reaches KEEP's Control Plane, see Security & Data Ownership.

What support to expect

Installation is designed to complete without KEEP's engineering team needing to intervene. (Source: KEEP Evaluation Program documentation)

Ending the evaluation

If you decide not to continue, a complete removal procedure exists for the software and data on your own infrastructure. (Source: KEEP Evaluation Program documentation) For the full removal process — including the one open question about your account record on KEEP's Control Plane — see Security & Data Ownership.

Next steps