Changelog

Initial public release of AgentLift. Define agents in agent.yaml, deploy with agentlift deploy, and get automatic tracing, cost tracking, and an OpenAI-compatible endpoint out of the box.

Added LLM-as-judge evaluation support. Configure eval criteria in agent.yaml and run agentlift eval to score your agent’s responses against defined quality benchmarks.

CLI improvements: faster agentlift init scaffolding, better error messages for invalid agent.yaml configs, and agentlift logs --follow for real-time trace streaming.