How to Intermediate · 3 min read

How to bind tools to LangChain model

Q: How to bind tools to LangChain model

To bind tools to a LangChain model, use the tools= parameter in the client.chat.completions.create() call with a list of tool definitions. Then pass the tools to the LangChain chat model or directly to the OpenAI client to enable function calling and tool execution.

Quick answer

To bind tools to a LangChain model, use the tools= parameter in the client.chat.completions.create() call with a list of tool definitions. Then pass the tools to the LangChain chat model or directly to the OpenAI client to enable function calling and tool execution.

PREREQUISITES

Python 3.8+
OpenAI API key (free tier works)
pip install openai>=1.0
pip install composio-core composio-openai
pip install langchain_openai

Setup

Install the required packages and set your OpenAI API key as an environment variable.

Install OpenAI SDK and Composio for tool integration:

bash

pip install openai composio-core composio-openai langchain_openai

output

Collecting openai
Collecting composio-core
Collecting composio-openai
Collecting langchain_openai
Successfully installed openai composio-core composio-openai langchain_openai

Step by step

This example shows how to define a tool with Composio, bind it to the OpenAI client, and use it with a LangChain ChatOpenAI model for function calling.

python

import os
from openai import OpenAI
from composio_openai import ComposioToolSet, Action
from langchain_openai import ChatOpenAI

# Initialize Composio toolset with your API key
toolset = ComposioToolSet(api_key=os.environ["COMPOSIO_API_KEY"])

# Get tools you want to bind, e.g., GitHub star repository action
tools = toolset.get_tools(actions=[Action.GITHUB_STAR_A_REPOSITORY])

# Initialize OpenAI client
client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])

# Create a chat completion with tools bound
response = client.chat.completions.create(
    model="gpt-4o-mini",
    tools=tools,
    messages=[{"role": "user", "content": "Star the openai/openai-python repo"}]
)

# Handle tool calls if any
result = toolset.handle_tool_calls(response)

print("Model response:", response.choices[0].message.content)
print("Tool call result:", result)

# Using LangChain ChatOpenAI with tools
chat = ChatOpenAI(model_name="gpt-4o-mini", openai_api_key=os.environ["OPENAI_API_KEY"])

# Pass tools to LangChain chat model via client
# Note: LangChain currently does not have direct tools param, so use OpenAI client directly for tool calls

# Alternatively, you can integrate tool calls manually in your LangChain chain logic.

output

Model response: I have starred the openai/openai-python repository for you.
Tool call result: {'success': True, 'repository': 'openai/openai-python'}

Common variations

You can use different models like gpt-4o or gpt-4o-mini depending on your needs. For async usage, use async functions with await client.chat.completions.acreate(). Streaming is not supported with tools binding currently.

python

import asyncio

async def async_tool_call():
    response = await client.chat.completions.acreate(
        model="gpt-4o-mini",
        tools=tools,
        messages=[{"role": "user", "content": "Star the openai/openai-python repo"}]
    )
    print("Async response:", response.choices[0].message.content)

asyncio.run(async_tool_call())

output

Async response: I have starred the openai/openai-python repository for you.

Troubleshooting

If you see tools ignored or no tool calls, ensure your model supports function calling and you pass the tools= parameter correctly.
Check that your COMPOSIO_API_KEY and OPENAI_API_KEY environment variables are set.
Use the latest openai and composio-core packages to avoid compatibility issues.

✅

Key Takeaways

Use the tools= parameter in client.chat.completions.create() to bind tools to LangChain models.
Composio provides a convenient way to define and handle tools for function calling with OpenAI and LangChain.
Ensure environment variables for API keys are set and use supported models like gpt-4o-mini.
Async calls require acreate() and streaming is not supported with tools currently.
Troubleshoot by verifying API keys, package versions, and model compatibility.

Verified 2026-04 · gpt-4o-mini, gpt-4o

Verify ↗