Compass LangChain Toolkit

The langchain-compass toolkit contains tools that enable an LLM agent to interact with popular DeFi protocols non-custodially i.e. tools return unsigned transactions. The toolkit is built on top of a Universal DeFi API (Compass API) allowing agents to perform financial operations like:

• Swapping tokens on Uniswap and Aerodrome
• Lending or borrowing assets using protocols on Aave
• Providing liquidity on Aerodrome and Uniswap
• Transferring funds between wallets.
• Querying balances, portfolios and monitoring positions.

Overview

Integration details

Class	Package	Serializable	JS support	Package latest
LangchainCompassToolkit	langchain-compass	❌	❌

Tool features

Here’s a sample of the tools this toolkit provides (subject to change daily):

• aave_supply_: Supply assets to Aave to earn interest.
• aave_borrow_: Borrow assets from Aave using collateral.
• uniswap_swap_sell_exactly_: Swap a specific amount of one token on Uniswap.
• generic_portfolio_get_: Retrieve a wallet’s portfolio in USD and token balances.
• generic_transfer_erc20_: Transfer ERC20 tokens between addresses.

Setup

Here we will:

1. Install the langchain package
2. Import and instantiate the toolkit
3. Pass the tools to your agent with toolkit.get_tools()

Installation

This toolkit lives in the langchain-compass package:

%pip install -qU langchain-compass

Environment Setup

To run these examples, ensure LangChain has access to an LLM service. For instance, if you're using GPT-4o, create a .env file containing:

# .env file
OPENAI_API_KEY = "your_openai_api_key_here"

Instantiation

Now we can instantiate our toolkit:

from langchain_compass.toolkits import LangchainCompassToolkit

toolkit = LangchainCompassToolkit(compass_api_key=None)

Tools

View available tools:

tools = toolkit.get_tools()
for tool in tools:
    print(tool.name)

Use within an agent

We will need a LLM or chat model:

from dotenv import load_dotenv
from langchain_openai import ChatOpenAI

load_dotenv()

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

API Reference:ChatOpenAI

Initialize the agent with the tools:

from langgraph.prebuilt import create_react_agent

tools = toolkit.get_tools()
agent_executor = create_react_agent(llm, tools)

API Reference:create_react_agent

Example usage:

example_query = "what is the balance of vitalic.eth."  # spelt wrong intentionally

events = agent_executor.stream(
    {"messages": [("user", example_query)]},
    stream_mode="values",
)
for event in events:
    event["messages"][-1].pretty_print()

Expected output:

The portfolio value of the wallet address associated with `vitalik.eth` is approximately $485,515.45 USD. Below is a detailed breakdown of the token balances:

1. **1INCH**: `6.04` tokens valued at approximately `$1.02` USD
2. **AAVE**: `0.0102` tokens valued at approximately `$1.40` USD
3. **BAL**: `0.9321` tokens valued at approximately `$0.96` USD
4. **crvUSD**: `0.7755` tokens valued at approximately `$0.78` USD
5. **DAI**: `317,203.87` tokens valued at approximately `$317,292.76` USD
6. **ENS**: `1,144.04` tokens valued at approximately `$16,329.73` USD
7. **LINK**: `1.78` tokens valued at approximately `$21.95` USD
8. **rsETH**: `0.00003` tokens valued at approximately `$0.05` USD
9. **UNI**: `0.000017` tokens valued at approximately `$0.00009` USD
10. **USDC**: `123,223.71` tokens valued at approximately `$123,215.08` USD
11. **USDT**: `170.15` tokens valued at approximately `$170.12` USD
12. **WBTC**: `0.00107` tokens valued at approximately `$90.97` USD
13. **WETH**: `16.40` tokens valued at approximately `$28,390.62` USD

The total value from all these assets comes to approximately $485,515.45 USD.

Related

• Tool conceptual guide
• Tool how-to guides