Documentation menu

Get Started

Jetty provides all the components to run and evaluate your AI workload. Under the hood, Jetty takes care of all the infrastructure to run an agent, track everything it does and record the outputs from a run. When the run completes, the results are organized along with a workload report that indicate how well the agent performed. No infrastructure to manage, no containers to configure, no agent lifecycle to babysit.

From sign-up to your first evaluated run

  1. Create an account. Sign up free. Nothing to install: pick a runbook from the directory and run it from your account.
  2. Connect your agent. Install the Jetty plugin in your coding agent, or connect your app through the MCP server, the Client SDK or the API.
  3. Author a runbook. One markdown file with the job, the bar for “done”, and the checks.
  4. Run your AI workload. Jetty runs it in a managed sandbox, records every step, and scores the result against the checks in the runbook.