Crypto Trading Agent

This project demonstrates how to use Composio to create an automated crypto trading agent that monitors and executes trades on Base Mainnet.

Overview

The Crypto Trading Agent is an AI Agent that monitors cryptocurrency prices and automatically executes trades between USDC and ETH when favorable conditions are met on Base Mainnet.

Getting Started

Python

1. Installation

   install dependencies

   pip install composio-langchain python-dotenv cdp-langchain langchain-openai
   
2. Connecting to tools and models

   connect to required tools

   composio add coinbase
   composio add exa
   
   export OPENAI_API_KEY="<your-openai-api-key>"
   
3. Importing the required libraries

   import required libraries

   from langchain_openai import ChatOpenAI
   from langchain.agents import create_openai_functions_agent, AgentExecutor
   from langchain import hub
   from composio_langchain import ComposioToolSet, App
   from cdp_langchain.agent_toolkits import CdpToolkit
   from cdp_langchain.utils import CdpAgentkitWrapper
   from dotenv import load_dotenv
   from pathlib import Path
   import time
   import os
   
   load_dotenv()
   
4. Setting up the Tools and Agent

   setup tools and agent

   llm = ChatOpenAI(model="gpt-4o")
   
   toolset = ComposioToolSet()
   tools = toolset.get_tools(apps=[
       App.COINBASE,
       App.EXA,
       App.FILETOOL
   ])
   
   prompt = hub.pull("hwchase17/openai-functions-agent")
   agent = create_openai_functions_agent(llm, tools, prompt)
   agent_executor = AgentExecutor(agent=agent, tools=tools, verbose=True)
   
5. Initializing Wallet

   initialize wallet

   # Find and store wallet data
   agent_executor.invoke({
       "input": 'Find existing Wallet on Base Mainnet (If not found, create a new one), Get its details including wallet_id, seed, and default_address_id, store it in the wallet_data.txt file in this format: {"wallet_id": "", "seed": "", "default_address_id": ""}'
   })
   
   # Get the current file's directory and read wallet data
   current_dir = Path(__file__).parent
   wallet_data_file = current_dir / "wallet_data.txt"
   
   try:
       with open(wallet_data_file) as f:
           wallet_data = f.read()
   except FileNotFoundError:
       print(f"Could not find wallet data file at: {wallet_data_file}")
       wallet_data = None
   
6. Setting up CDP Toolkit

   setup cdp toolkit

   os.environ['NETWORK_ID'] = 'base-mainnet'
   
   # Configure CDP Agentkit
   values = {}
   if wallet_data is not None:
       values = {"cdp_wallet_data": wallet_data}
   
   cdp = CdpAgentkitWrapper()
   toolkit = CdpToolkit.from_cdp_agentkit_wrapper(cdp)
   cdp_toolkit = toolkit.get_tools()
   cdp_toolkit.extend(tools)
   
   agent = create_openai_functions_agent(llm, cdp_toolkit, prompt)
   agent_executor = AgentExecutor(agent=agent, tools=cdp_toolkit, verbose=True)
   
7. Running the Trading Loop

   run trading loop

   POLLING_INTERVAL = 60  # Poll every 60 seconds
   
   print("Starting trading poll...")
   while True:
       try:
           print(f"Executing trade at {time.strftime('%Y-%m-%d %H:%M:%S')}")
           agent_executor.invoke({
               "input": "Check the prices of USDC and ETH on Base Mainnet and perform a trade between 0.01 ETH and USDC or 0.01 USDC and ETH if the price is favorable on Base Mainnet"
           })
       except Exception as e:
           print(f"Error during trade execution: {e}")
       
       print(f"Waiting {POLLING_INTERVAL} seconds before next trade...")
       time.sleep(POLLING_INTERVAL)
   
8. Final Code

   final code

   from langchain_openai import ChatOpenAI
   from langchain.agents import create_openai_functions_agent, AgentExecutor
   from langchain import hub
   from composio_langchain import ComposioToolSet, App
   from cdp_langchain.agent_toolkits import CdpToolkit
   from cdp_langchain.utils import CdpAgentkitWrapper
   from dotenv import load_dotenv
   from pathlib import Path
   import time
   import os
   
   load_dotenv()
   
   llm = ChatOpenAI(model="gpt-4o")
   
   toolset = ComposioToolSet()
   tools = toolset.get_tools(apps=[
       App.COINBASE,
       App.EXA,
       App.FILETOOL
   ])
   
   prompt = hub.pull("hwchase17/openai-functions-agent")
   agent = create_openai_functions_agent(llm, tools, prompt)
   agent_executor = AgentExecutor(agent=agent, tools=tools, verbose=True)
   
   agent_executor.invoke({
       "input": 'Find existing Wallet on Base Mainnet, Get its details including wallet_id, seed, and default_address_id, store it in the wallet_data.txt file in this format: {"wallet_id": "", "seed": "", "default_address_id": ""}'
   })
   
   current_dir = Path(__file__).parent
   wallet_data_file = current_dir / "wallet_data.txt"
   
   try:
       with open(wallet_data_file) as f:
           wallet_data = f.read()
   except FileNotFoundError:
       print(f"Could not find wallet data file at: {wallet_data_file}")
       wallet_data = None
   
   os.environ['NETWORK_ID'] = 'base-mainnet'
   
   values = {}
   if wallet_data is not None:
       values = {"cdp_wallet_data": wallet_data}
   
   cdp = CdpAgentkitWrapper()
   toolkit = CdpToolkit.from_cdp_agentkit_wrapper(cdp)
   cdp_toolkit = toolkit.get_tools()
   cdp_toolkit.extend(tools)
   
   agent = create_openai_functions_agent(llm, cdp_toolkit, prompt)
   agent_executor = AgentExecutor(agent=agent, tools=cdp_toolkit, verbose=True)
   
   POLLING_INTERVAL = 60  # Poll every 60 seconds
   
   print("Starting trading poll...")
   while True:
       try:
           print(f"Executing trade at {time.strftime('%Y-%m-%d %H:%M:%S')}")
           agent_executor.invoke({
               "input": "Check the prices of USDC and ETH on Base Mainnet and perform a trade between 0.01 ETH and USDC or 0.01 USDC and ETH if the price is favorable on Base Mainnet"
           })
       except Exception as e:
           print(f"Error during trade execution: {e}")
       
       print(f"Waiting {POLLING_INTERVAL} seconds before next trade...")
       time.sleep(POLLING_INTERVAL)