Router

Route prompts to different agents.

Constructor

Router(
    agents: dict[str, Agent],
    route: Callable[[str], str] | None = None,
    model: str | None = None,
    provider: str = "mistral",
    api_key: str | None = None,
)

Parameter	Type	Description
agents	dict[str, Agent]	Named agents to route between
route	Callable	Function (prompt) -> agent_name. If None, uses LLM
model	str	Model for LLM-based routing
provider	str	Provider for LLM-based routing
api_key	str	API key for LLM-based routing

Methods

route

async def route(prompt: str) -> str

Determine which agent to use. Returns agent name.

run

async def run(prompt: str, plan: bool = False) -> str

Route to appropriate agent and run it.

Parameter	Type	Description
prompt	str	The prompt to process
plan	bool	If True, agent creates plan first

run_sync

def run_sync(prompt: str, plan: bool = False) -> str

Synchronous version of run().

Example

from pure_agents import Agent, Router

coder = Agent(template="coder")
writer = Agent(template="creative")

# Function-based routing
def route(prompt: str) -> str:
    if "code" in prompt.lower():
        return "coder"
    return "writer"

router = Router(
    agents={"coder": coder, "writer": writer},
    route=route,
)

result = await router.run("Write a sorting function")

LLM-based routing

# No route function = LLM decides
router = Router(
    agents={"coder": coder, "writer": writer},
)

result = await router.run("Explain recursion")