Skip to main content

Specification driven evaluation workflows

Spec-driven infrastructure to automate validation and monitoring for every agent you build or buy. Reclaim hundreds of engineering hours in ad-hoc testing and build launch confidence with objective, machine-readable rigor

Get Started

Start with a guided walkthrough, build a project from scratch, or learn the asset model that connects projects, specifications, evals, and results.

Onboarding Example

Use the seeded demo flow to understand how Projects, Datasets, Agents, Specs, Evals, and Results fit together.

Open section

Quickstart

Build your own setup from scratch: create a project, add assets, run an eval, and review the output.

Open section

Core Concepts

Learn the reusable evaluation model behind Gold Team, Red Team, Specifications, Evals, Playground, and Results.

Open section