OpenAI Agents

OpenAI Agents is a lightweight framework for creating agentic workflows using agent swarms, handoffs, and tool use. Confident AI also allows you to trace OpenAI Agent workflows with one line of code.

Quickstart

Install deepeval:


pip install -U deepeval


deepeval login --confident-api-key YOUR_API_KEY

Trace your OpenAI Agent workflows:


from agents import Agent, Runner, add_trace_processor
from deepeval.openai_agents import DeepEvalTracingProcessor
 
add_trace_processor(DeepEvalTracingProcessor())
 
# Replace with your agent code
agent = Agent(name="Assistant", instructions="You are a helpful assistant")
result = Runner.run_sync(agent, "Write a haiku about recursion in programming.")
print(result.final_output)

DeepEvalTracingProcessor handles both asynchronous workflows and streamed responses.

Loading video...

Tracing OpenAI Agent Example

0 views • 0 days ago

Confident AI

100K subscribers

Confident AI’s DeepEvalTracingProcessor automatically traces the following span types:

AgentSpanData: OpenAI span for agents
FunctionSpanData: OpenAI span for function calls
MCPListToolsSpanData: OpenAI span for MCP tool calls
GenerationSpanData: OpenAI span for LLM chat completions
ResponseSpanData: OpenAI span for LLM response creations
HandoffSpanData: OpenAI span for agent handoffs
GuardrailSpanData: OpenAI span for guardrails triggered

Logging Threads

Chat conversation threads are also automatically logged to Confident AI when a group_id is provided in your OpenAI Agent workflow.


from agents import Agent, Runner, trace, add_trace_processor
from deepeval.openai_agents import DeepEvalTracingProcessor
 
add_trace_processor(DeepEvalTracingProcessor())
 
thread_id = "unique_thread_id"
 
async def main():
    agent = Agent(name="Assistant", instructions="Reply very concisely.")
 
    with trace(workflow_name="Conversation", group_id=thread_id):
        # First turn
        result = await Runner.run(agent, "What city is the Golden Gate Bridge in?")
        print(result.final_output)
        # San Francisco
 
        # Second turn
        new_input = result.to_input_list() + [{"role": "user", "content": "What state is it in?"}]
        result = await Runner.run(agent, new_input)
        print(result.final_output)
        # California