1
Build your agent
Create an agent by connecting Prompt, Code, and API nodes based on your data flow. Each node type handles a specific task in your AI workflow.
2
Configure your test
Test your agent against a Dataset and add Evaluators to measure the quality of outputs. Configure any additional parameters needed for your test run.
3
Review results
Analyze the test report for quality metrics like accuracy and performance metrics like latency and cost. Use these insights to iterate on your agent.