Streaming graph events

Overview

• Graph streaming exposes progress such as state updates, node events, and model token events.
• It lets a UI distinguish “retrieving,” “waiting for approval,” and “complete.” Stream modes and event APIs vary by version.

Why this matters

A graph may perform several steps before producing final text.

An everyday example

A delivery tracker shows “picked up,” “in transit,” and “delivered.”

A technical example

Translate an internal retrievePolicy completion into a public progress: "Found policy information" event.

How it works

1. Subscribe server-side
2. Allowlist public event fields
3. Remove sensitive state
4. Emit stable application events
5. Finish with final status.

Architecture and flow

Graph events → event projector/redactor → SSE → UI.

The insurance assistant example

The customer sees “Waiting for your confirmation,” while staff-only notes stay private.

When to use it

Interactive multi-step tasks and visible progress.

When to avoid it

Never forward raw state snapshots or all tool results to the browser.

Interview answer

“I stream application-safe progress events rather than raw graph state, with completion, pause, and error events explicitly modeled.”

Benefits, trade-offs, and production considerations

• An internal/private state field is not automatically hidden in all streaming modes.
• Apply a server-side public event projection regardless of graph schemas.
• Test reconnect and event ordering.
• LangGraph streaming.