Back to library
AI
tutorial

Eliminate UI Lag with Managed Agent Event Deltas

Enable event deltas to stream agent responses character-by-character.

Use the 'event_deltas[]' query parameter in Managed Agents session streams to provide real-time, incremental previews of agent messages.

Claude

The Scenario

You are building a customer support dashboard where users interact with a Claude Managed Agent to solve technical issues. You want the user to see the agent's response as it's being typed to prevent them from thinking the system is frozen.

Before & after

The old way

Users previously had to wait for the entire agent message to process before seeing any output. This caused a perceived delay of 10–30 seconds depending on response length.

With AI

Developers can now enable 'event_deltas' to fetch character-by-character previews. Setting this up takes roughly 2 minutes of configuration.

The Prompt

I am implementing Claude Managed Agents. Provide a Python snippet using the 'managed-agents-2026-04-01' beta header that calls 'GET /v1/sessions/[SESSION_ID]/events/stream' with the 'event_deltas[]' parameter to stream real-time message previews.

By using the 'event_deltas[]' query parameter on the session stream endpoint, you can render text incrementally. This significantly improves the perceived speed of your AI application by showing the agent's 'thinking' or 'writing' in real-time.

Source

Claude Platform - Claude Platform Docs
"event_start and event_delta events preview an agent message's text as it's generated, before the complete agent.message event arrives."