Beginner
Your first evaluation
Install the SDK, create one run, inspect the response. The code path is the one new teams need first.
Quickstarts first. Deeper rollout paths next. Every walkthrough ends with a working artifact you can keep, replay, and adapt — repo-level eval material and reviewed exports, not a sandbox you throw away.
First run, regression bank, field workflow setup, and reviewed rollout. Each one ends with a working artifact you can replay and adapt.
Code paths follow the SDK. When the API shape moves, the walkthrough moves with it. No stale snippets.
Install, run, then review. You leave with an export a reviewer can open, not a sandbox you throw away.
Each card shows the level, the duration, and the kind of artifact you should leave with. No diagrams. Working code.
Beginner
Install the SDK, create one run, inspect the response. The code path is the one new teams need first.
Intermediate
Turn caught failures into replayable tests. Run the bank before every release. Catch what already slipped once.
Intermediate
Route run.completed, export.ready, and escalation events into CI, incident response, and the archive.
Advanced
Start from a real input in your field, route the hard cases, and define the signed export the workflow has to produce.
Intermediate
Move from a working evaluation to release gates, escalation paths, and exports that survive an audit.
Intermediate
Open the signed proof packet, walk the rubric history, and answer the reviewer in one window.
The walkthrough finishes when something is running on your own infrastructure. You keep the artifact. The next page tells you what to wire next.
An install command, a clean repo, and a question your team can name.
A working artifact, a reviewed reference, and the next step already named.