Loading connector details…
Loading connector details…
Choose a unique username to continue using AgentHotspot
by iris-eval • Uncategorized
An open-source MCP server that evaluates AI agent outputs for quality, safety, and cost.
Automatic evaluation of output quality and safety without code changes
Detailed cost tracking and budget enforcement
Real-time monitoring and trace visualization via a web dashboard
Iris is an open-source MCP server designed to score output quality, detect safety failures, and enforce cost budgets for AI agents automatically without requiring SDKs or code changes. It provides detailed trace logging, built-in evaluation rules, cost visibility, and a real-time web dashboard for monitoring. Iris supports multiple transport methods and offers extensibility with custom evaluation rules, making it ideal for production environments to ensure agent reliability and cost control.