01
What exists?
Bring resources, relationships, and changes into one view instead of piecing them together across tools.
IIP / Learning preview available / Production v1 pending
Infrastructure intelligence, grounded in evidence.
IIP helps operators understand what they run, what changed, and why it matters. Its AI FinOps capability adds a practical question: where is our AI money going, and what can we do about it?
Five questions. One connected picture.
02
What happened?
Investigate a bounded question and follow the evidence behind each conclusion.
03
Where is AI spend going?
Connect model usage to estimated cost, then attribute it to reviewed applications and teams.
04
What changed?
Compare usage and cost over defined windows, with missing data and coverage gaps made visible.
05
What could we improve?
Review rule-based savings opportunities linked to evidence—not a promise of automatic savings.
Privacy and control are part of the design.
Metadata-only collection by default. No inference proxy, no prompts or responses by default, and asynchronous telemetry designed to fail open for the application. Tenant scope, bounded investigations, and explicit approval for governed actions remain separate controls.
A learning release, not production v1.
The Apache-2.0 learning preview is public now. Build it locally with Docker to explore synthetic AI usage, fixture-priced cost, and a read-only resource investigation—without AWS, Kubernetes, provider keys, or a host Python installation. Keep real applications and customer data out of this disposable demo.
Signed production artifacts, live-provider qualification, supported deployment and lifecycle evidence, and public v1 readiness remain separate work. The persistent community preview is not an automatic upgrade from the learning environment.
Infrastructure Intelligence Platform and IIP are neutral working names. The DevOps Human is the site and GitHub owner, not a finalized product brand.