LangChain Adds Google ADK Agent Deployment to LangSmith

LangChainLangChain

LangChain now enables developers to deploy Google Agent Development Kit (ADK) agents directly to LangSmith. This integration provides built-in session persistence, streaming, and tracing for ADK agents on LangSmith's managed infrastructure. It simplifies moving ADK-built agent prototypes into production environments.

LangChain has integrated support for deploying Google Agent Development Kit (ADK) agents to its LangSmith Agent Server. This is achieved through the deployments-wrap-sdk package, which provides a wrapper to convert a configured ADK Runner into a LangGraph-compatible graph, streamlining the deployment process.
Wrapper Package
deployments-wrap-sdk
Required Python Version
3.11+
Required CLI
LangGraph CLI
Supported Agent Primitives
Agent, SequentialAgent, ParallelAgent
Supported Tools
Python function tools, LongRunningFunctionTool
Supported Models
Gemini models, LiteLLM adapter models

This update bridges Google's agent development framework with LangChain's platform for managing the agent lifecycle. It addresses the need for production-grade features by offering session persistence, real-time streaming of token events, and automatic tracing for observability, which are critical for reliable agent operations.

Deployment of ADK agents to LangSmith is possible with a single function call, requiring Python 3.11+, the LangGraph CLI, and a LangSmith API key. The integration supports ADK agent primitives, Python function tools, and models like Gemini directly, or others via ADK’s LiteLLM adapter.

View the full update on docs.langchain.com
LangChain
LangChain
@LangChain
X

You can now deploy Google ADK agents to LangSmith! Wrap your ADK agent with one function call and deploy it to managed infra, with built in: ✅ Session persistence ✅ Streaming ✅ Tracing https://t.co/QDzGlOJoQg

4retweets29likes
View on X

Still wondering? A few quick answers below.

LangChain has enabled the deployment of Google Agent Development Kit (ADK) agents to its LangSmith Agent Server. This allows developers to take ADK agents and deploy them to LangSmith's managed infrastructure using a wrapper function.

When deployed to LangSmith, ADK agents gain built-in session persistence, ensuring agent state survives restarts. They also get token streaming for real-time output and automatic LangSmith tracing for detailed observability of agent execution.

To deploy ADK agents to LangSmith, you need Python 3.11 or newer, the LangGraph CLI for local development and deployment, and a LangSmith API key. A Google AI API key is also required if you are using Gemini models.

The wrapper does not support multimodal input (only text is forwarded), multiple new messages per turn, bidirectional/live streaming, non-text output parts, or intermediate events as separate messages. It also requires LangsmithSessionService for session persistence.

Every HeadsUpAI update is written based on its original source and reviewed before it's published. Read our editorial standards →

Share this update