AgentOpt checks AI agents before you deploy them: finds structural problems in an agent workflow, simulates cheaper configurations, audits a process document against best practices, and reports what a Claude Code session actually used (turns, tool calls, tokens, cost, tool success rate).
This page renders its API reference with JavaScript. Without it, read /llms.txt for a machine-readable summary, /guide for the tester walkthrough, or /openapi.json for the raw schema.