Build distributed agents that spawn more distributed sub-agents.
nFactorial is a distributed task queue for building reliable multi-agent systems. It makes the following trivial to implement:
- Agent Reliability: Automatic retries, backoff strategies, and recovery of dropped tasks from crashed workers.
- In-flight Task Management: Cancel, steer, and monitor running tasks.
- Spawning Sub Agents: Having an agent spawn multiple sub agents and wait for their completion before continuing.
- Deferred Tools: Pause the agent while it waits for long running tools to complete externally or wait for user approval before continuing.
- Observability: Built-in metrics dashboard and comprehensive logging
pip install nfactorialfrom factorial import Agent, Orchestrator, AgentWorkerConfig, gpt_41
def get_weather(location: str) -> str:
return f"The weather in {location} is sunny and 72°F"
agent = Agent(
instructions="You help users get weather information.",
model=gpt_41,
tools=[get_weather],
)
# Create orchestrator
orchestrator = Orchestrator(
redis_host="localhost",
redis_port=6379,
redis_db=0,
redis_max_connections=50,
)
orchestrator.register_runner(
agent=agent, agent_worker_config=AgentWorkerConfig(workers=1)
)
# Run the system
if __name__ == "__main__":
orchestrator.run()/docs/examples/code_agent
/docs/examples/multi_agent
