Router¶

Agent router for multi-agent routing and fallback.

class agent.router.AgentRouter(agents, strategy=RoutingStrategy.FALLBACK, custom_router=None)[source]¶

Bases: object

Routes requests across multiple agents with fallback support.

The router provides strategies for: - Automatic failover when providers fail - Load balancing across providers - Capability-based routing - Cost optimization

Example

```python router = AgentRouter(

agents=[
Agent(provider=”anthropic”, model=”claude-sonnet”), Agent(provider=”openai”, model=”gpt-4o”),

], strategy=”fallback”,

)

# Automatically falls back if first provider fails response = router.run(“Hello!”) ```

Parameters:

agents (list[Agent])
strategy (RoutingStrategy | str)
custom_router (Callable[[AgentRequest, list[Agent]], RouteResult] | None)

run(input=None, *, messages=None, system=None, temperature=None, max_tokens=None, stop=None, metadata=None)[source]¶

Execute a request using the routing strategy.

Parameters:

input (str | None) – User input text
messages (list[Message] | None) – Optional message history
system (str | None) – System prompt
temperature (float | None) – Sampling temperature
max_tokens (int | None) – Max tokens
stop (list[str] | None) – Stop sequences
metadata (dict[str, Any] | None) – Request metadata

Return type:

AgentResponse

Returns:

AgentResponse from the selected agent

Raises:

RoutingError – If all agents fail

async run_async(input=None, *, messages=None, system=None, temperature=None, max_tokens=None, stop=None, metadata=None)[source]¶

Execute a request asynchronously using the routing strategy.

Parameters:

input (str | None) – User input text
messages (list[Message] | None) – Optional message history
system (str | None) – System prompt
temperature (float | None) – Sampling temperature
max_tokens (int | None) – Max tokens
stop (list[str] | None) – Stop sequences
metadata (dict[str, Any] | None) – Request metadata

Return type:

AgentResponse

Returns:

AgentResponse from the selected agent

stream(input=None, *, messages=None, system=None, temperature=None, max_tokens=None, metadata=None)[source]¶

Stream a response using the routing strategy.

For fallback strategy, only tries next agent if initial connection fails.

Parameters:

input (str | None) – User input text
messages (list[Message] | None) – Optional message history
system (str | None) – System prompt
temperature (float | None) – Sampling temperature
max_tokens (int | None) – Max tokens
metadata (dict[str, Any] | None) – Request metadata

Return type:

StreamResponse

Returns:

StreamResponse from the selected agent

async stream_async(input=None, *, messages=None, system=None, temperature=None, max_tokens=None, metadata=None)[source]¶

Stream a response asynchronously using the routing strategy.

Parameters:

input (str | None) – User input text
messages (list[Message] | None) – Optional message history
system (str | None) – System prompt
temperature (float | None) – Sampling temperature
max_tokens (int | None) – Max tokens
metadata (dict[str, Any] | None) – Request metadata

Return type:

AsyncStreamResponse

Returns:

AsyncStreamResponse from the selected agent

json(input=None, *, schema, messages=None, system=None, temperature=None, max_tokens=None, metadata=None)[source]¶

Execute a structured output request with routing.

Parameters:

input (str | None) – User input text
schema (type[BaseModel] | dict[str, Any]) – Output schema
messages (list[Message] | None) – Optional message history
system (str | None) – System prompt
temperature (float | None) – Sampling temperature
max_tokens (int | None) – Max tokens
metadata (dict[str, Any] | None) – Request metadata

Return type:

AgentResponse

Returns:

AgentResponse with parsed output