Runtime Configuration

Supported in ADKPython v0.1.0TypeScript v0.2.0Go v0.1.0Java v0.1.0Kotlin v0.1.0

RunConfig controls how agents behave at runtime, including streaming mode, speech settings, LLM call limits, and live agent options. Pass a RunConfig to runner.run_async() or runner.run_live() to override default behavior.

PythonTypeScriptGoJavaKotlin

from google.adk.agents.run_config import RunConfig, StreamingMode

config = RunConfig(
    streaming_mode=StreamingMode.SSE,
    max_llm_calls=200,
)

async for event in runner.run_async(
    ...,
    run_config=config,
):
    ...

import { RunConfig, StreamingMode } from '@google/adk';

const config: RunConfig = {
  streamingMode: StreamingMode.SSE,
  maxLlmCalls: 200,
};

import "google.golang.org/adk/v2/agent"

config := agent.RunConfig{
    StreamingMode: agent.StreamingModeSSE,
}

import com.google.adk.agents.RunConfig;
import com.google.adk.agents.RunConfig.StreamingMode;

RunConfig config = RunConfig.builder()
    .streamingMode(StreamingMode.SSE)
    .maxLlmCalls(200)
    .build();

val config =
    RunConfig(
        streamingMode = StreamingMode.SSE,
        // Cap the LLM calls a single run may make. Defaults to 500.
        maxLlmCalls = 200,
    )

// Pass it to runner.runAsync
// runner.runAsync(..., runConfig = config)

Manage sessions and context

Supported in ADKPython

For long-running sessions, you can control how much history is loaded and whether the context window is compressed:

• get_session_config: Limits which events are fetched when loading a session. Use num_recent_events or after_timestamp to avoid loading the full event history on every invocation. These filters limit the loaded view without deleting stored events. New events are appended to the stored history, preserving older events excluded from the loaded view.
• context_window_compression: Enables context window compression for LLM input, useful when sessions approach model context limits.
• include_thoughts_from_other_agents: Controls whether thought parts from other agents are included in the LLM context. Disabled by default.