Use this pattern when later runs need continuity from earlier runs.

• multi-step product flows
• coding copilots with follow-up questions
• iterative planning and refinement
• user-specific assistant sessions

from loom import Agent, Model, RunContext, Runtime, SessionConfig

agent = Agent(
    model=Model.anthropic("claude-sonnet-4"),
    instructions="You help users refine API designs across multiple turns.",
    runtime=Runtime.sdk(),
)

session = agent.session(SessionConfig(id="user-123"))

first = await session.run("List three options for versioning an API")
second = await session.run(
    "Recommend one option for a startup team",
    context=RunContext(inputs={"previous_answer": first.output}),
)

Use Session for continuity and RunContext for explicit per-run inputs.

That split keeps the contract clear:

• session state is long-lived
• RunContext is request-scoped

• add SessionRestorePolicy when reopening a durable session
• add FileSessionStore when state must survive process restart
• add KnowledgeBundle in RunContext when grounding is needed on one run

• examples/04_multi_task_session.py
• examples/03_events_and_artifacts.py