Multi-Agent API

---

Graph

from elsai.multiagent import Graph

Deterministic, dependency-driven agent orchestration.

Graph.__init__

Graph(
    *,
    graph_id=None,
    state=None,
    hooks=None,
    plugins=None,
    session_manager=None,
    trace_attributes=None,
    max_iterations=None,
)

Graph.add_node

graph.add_node(agent)

Add an Agent as a node in the graph.

Graph.add_edge

graph.add_edge(source_agent, target_agent)

Create a directed edge so target_agent runs after source_agent and receives its output.

Graph.__call__

result = graph(prompt)

Execute the graph synchronously.

Graph.invoke_async

result = await graph.invoke_async(prompt)

Execute the graph asynchronously.

GraphResult

result.output          # str: final combined output
result.status          # "completed" | "failed" | "interrupted"
result.node_results    # dict[str, NodeResult]: per-node results
result.metrics         # Metrics: total token usage
result.execution_time  # float: total time in ms

---

Swarm

from elsai.multiagent import Swarm

Autonomous, collaborative agent orchestration.

Swarm.__init__

Swarm(
    agents,
    *,
    swarm_id=None,
    state=None,
    hooks=None,
    plugins=None,
    session_manager=None,
    trace_attributes=None,
)

Parameter	Type	Description
agents	list[Agent]	The agents in the swarm
swarm_id	str | None	Swarm identifier
state	AgentState | dict | None	Shared state
hooks	list | None	Lifecycle hooks

Swarm.__call__

result = swarm(prompt)

Swarm.invoke_async

result = await swarm.invoke_async(prompt)

SwarmResult

Same shape as GraphResult.

---

Agent.as_tool

tool = agent.as_tool(
    *,
    name=None,
    description=None,
    preserve_context=False,
)

Wraps the agent as an AgentTool that can be passed to another agent's tools parameter.

Parameter	Default	Description
name	agent.name	Tool name shown to the model
description	agent.description	Tool description
preserve_context	False	Preserve conversation history between calls

---

Multi-agent hooks

from elsai.hooks import (
    BeforeNodeCallEvent,
    AfterNodeCallEvent,
    MultiAgentInitializedEvent,
    BeforeMultiAgentInvocationEvent,
    AfterMultiAgentInvocationEvent,
)

BeforeNodeCallEvent

Fires before each agent (node) starts executing.

Attribute	Type	Description
node	Agent	The agent about to execute
input	str	Input being passed to the agent

AfterNodeCallEvent

Fires after each agent (node) finishes.

Attribute	Type	Description
node	Agent	The agent that just executed
result	NodeResult	The result from this node