Skip to main content

Runner

The Runner class executes agents and manages their lifecycle.

Import​

from openstackai import Runner

Basic Usage​

from openstackai import Agent, Runner

agent = Agent(name="Assistant", instructions="You are helpful")

# Sync execution
result = Runner.run_sync(agent, "Hello!")

# Async execution
result = await Runner.run(agent, "Hello!")

Methods​

run_sync​

result = Runner.run_sync(
    agent,
    messages,
    context=None,
    max_turns=10
)
ParameterTypeDefaultDescription
agentAgentrequiredAgent to run
messagesstr/listrequiredInput messages
contextdictNoneAdditional context
max_turnsint10Maximum conversation turns

run (async)​

result = await Runner.run(
    agent,
    messages,
    context=None,
    max_turns=10
)

run_stream​

async for event in Runner.run_stream(agent, messages):
    print(event)

Examples​

Simple Execution​

from openstackai import Agent, Runner

agent = Agent(
    name="Math Helper",
    instructions="You help with math problems"
)

# Run synchronously
result = Runner.run_sync(agent, "What is 25 * 4?")
print(result.final_output)  # "100"

With Context​

result = Runner.run_sync(
    agent,
    "Summarize the document",
    context={"document": "Long text here..."}
)

Streaming​

import asyncio
from openstackai import Agent, Runner

async def main():
    agent = Agent(name="Writer", instructions="You write stories")
    
    async for event in Runner.run_stream(agent, "Write a short story"):
        if hasattr(event, 'content'):
            print(event.content, end="", flush=True)

asyncio.run(main())

With Tools​

from openstackai import Agent, Runner, tool

@tool
def calculate(expression: str) -> str:
    """Calculate a math expression"""
    return str(eval(expression))

agent = Agent(
    name="Calculator",
    instructions="Use the calculate tool for math",
    tools=[calculate]
)

result = Runner.run_sync(agent, "What is 123 * 456?")

Multi-turn Conversation​

# Conversation with history
messages = [
    {"role": "user", "content": "My name is Alice"},
    {"role": "assistant", "content": "Hi Alice!"},
    {"role": "user", "content": "What's my name?"}
]

result = Runner.run_sync(agent, messages)
# Will remember context

Result Object​

result = Runner.run_sync(agent, "Hello")

print(result.final_output)    # Final response
print(result.messages)        # All messages
print(result.tool_calls)      # Tool calls made
print(result.tokens_used)     # Token consumption
print(result.cost)            # Estimated cost

Error Handling​

from openstackai import Runner, RunnerError

try:
    result = Runner.run_sync(agent, "Hello")
except RunnerError as e:
    print(f"Runner error: {e}")

See Also​

  • [[Agent]] - Agent class
  • [[Memory]] - Conversation memory
  • [[Workflows]] - Multi-agent workflows