pip install trajectoryevalMonitor, evaluate, and optimize your AI agents with comprehensive observability and intelligent evaluation frameworks. Ship with confidence.
From development to production, we've got you covered with enterprise-grade observability and evaluation tools.
Complete visibility into your AI agent's execution flow with detailed span-level insights.
Automated scoring with trajectory-level and span-level verifiers powered by LLM judges.
Identify bottlenecks and optimize your agent's performance with actionable insights.
Ensure your agents meet quality standards with comprehensive evaluation frameworks.
Integrate trajectory.ai into your existing AI agent workflow with just a few lines of code.
import trajectory
# Initialize the tracer
tracer = trajectory.init(
api_key="your-api-key",
project="my-ai-agent"
)
# Trace your agent execution
@tracer.trace("agent_execution")
def run_agent(query):
# Your agent logic here
result = agent.process(query)
return result
# Automatic evaluation
@tracer.evaluate([
trajectory.ToolOrderScorer(),
trajectory.LLMJudge(rubric="helpfulness")
])
def evaluate_agent_run(trace_id):
return tracer.get_trace(trace_id)Join thousands of developers who trust trajectory.ai to monitor and optimize their AI agents.