ADK Agent, Runner, and Session structure
Core components for building real-time voice AI agents with Google ADK.
# Describe an Agent agent = Agent( model=..., personality=..., tools=... ) # Hand it to a Runner runner = Runner(agent=agent) # Keep the Session in memory session = runner.create_session()
Language
python
Explanation
This snippet outlines the fundamental structure for creating a real-time voice AI agent using the Google ADK. It involves defining the agent's characteristics (model, personality, tools), initializing a runner to manage the live call, and maintaining the session state in memory.
What it does and when to use it
This Python snippet shows the basic ADK structure: defining an Agent with its model, personality, and tools, then passing it to a Runner to manage a live session, keeping the session state in memory.
Sources and credit
- Build a real-time voice AI agent with Google ADK and Gemini Live API ↗Google Cloud Tech · Google Cloud Tech on YouTube · Sep 14, 2026 · quoted as shared
First seen on Subvolts Sep 14, 2026. Updated Sep 15, 2026. Spotted a problem? Tell us.
Where it shows up
- Build a real-time voice AI agent with Google ADK and Gemini Live APIGoogle · Google Cloud Tech on YouTube · Sep 14, 2026
Related in the PlaybookPlaybook →
The Playbook: reusable how-tos, prompts, configs and tips pulled out of the news.
- Code snippetAgent Development Kit tool calling structure and callbackDemonstrates how to define tools and implement a callback to intercept tool calls before execution in Gemini agents. · 15 min
- ToolGoogle Gemini for real-time voice AI agents with tool callingEnables voice AI agents to take actions mid-conversation without breaking the flow using the Agent Development Kit. · 30 min
- ToolGoogle Codelabs for ADK instructionsInteractive tutorials for learning and using Google's Agent Development Kit (ADK). · 60 min
- Anti-patternBuilding voice AI like a simple API request/responseAvoid treating real-time voice AI as a single send-wait-get interaction; it requires continuous system management. · 15 min
- Code snippetInstrumenting AI agent code with Sigil SDKAdd lightweight wrappers to agent code using the Grafana Sigil SDK for observability. · 10 min
- ChecklistKey AI agent skill patterns for modular developmentSeven patterns for building effective, modular AI coding agents using skills. · 10 min
Questions people ask
- Which AI tools does this work with?
- The sources used it with Gemini.
- Where did this come from?
- Google Cloud Tech (Google Cloud Tech on YouTube). Subvolts collected, normalized and credited it; the link above opens the original.