Cdp agentkit toolkit integration

The Coinbase AgentKit LangChain extension gives an LLM agent tools to interact with the Coinbase Developer Platform. Agents can perform onchain operations such as transfers, trades, and smart contract interactions.

Overview

Integration details

Class / helper	Package	Serializable	JS support	Version
get_langchain_tools	coinbase-agentkit-langchain	❌	❌

Setup

1. Install the AgentKit packages.
2. Set CDP API credentials.
3. Create an AgentKit instance and convert it to LangChain tools with get_langchain_tools.
4. Pass the tools to your agent.

To enable automated tracing of individual tools, set your LangSmith API key:

import getpass
import os

os.environ["LANGSMITH_API_KEY"] = getpass.getpass("Enter your LangSmith API key: ")
os.environ["LANGSMITH_TRACING"] = "true"

Installation

pip install -qU coinbase-agentkit coinbase-agentkit-langchain

Set environment variables

Create API credentials in the CDP Portal, then set:

import getpass
import os

for env_var in [
    "CDP_API_KEY_ID",
    "CDP_API_KEY_PRIVATE",
]:
    if not os.getenv(env_var):
        os.environ[env_var] = getpass.getpass(f"Enter your {env_var}: ")

Instantiation

from coinbase_agentkit import AgentKit
from coinbase_agentkit_langchain import get_langchain_tools

agent_kit = AgentKit()
tools = get_langchain_tools(agent_kit)

For wallet providers, action providers, and advanced configuration, see the Coinbase AgentKit README.

Tools

for tool in tools:
    print(tool.name)

Available actions depend on the action providers configured on AgentKit. Common providers include wallet details, balances, faucet funds, transfers, trades, and token or NFT deployment.

Use within an agent

from langchain.agents import create_agent
from langchain_openai import ChatOpenAI

llm = ChatOpenAI(model="gpt-4o-mini")

agent = create_agent(model=llm, tools=tools)

For a full runnable example with a CDP smart wallet, see the AgentKit LangChain chatbot example.

API reference

For detailed documentation of configuration options and APIs:

• API reference
• AgentKit LangChain extension
• Source code