Skip to content

🦜️🧰 langchain-dev-utils

🚀 An efficient utility library crafted for LangChain and LangGraph developers

GitHub Repo PyPI Python Version License: MIT Last Commit

Why choose langchain-dev-utils?

Tired of writing repetitive code in LangChain development? langchain-dev-utils is exactly the solution you need! This lightweight yet powerful utility library is designed to enhance the LangChain and LangGraph development experience, helping you to:

  • Boost Development Efficiency - Reduce boilerplate code, allowing you to focus on core functionality.
  • Simplify Complex Workflows - Easily manage multi-model, multi-tool, and multi-agent applications.
  • Enhance Code Quality - Improve consistency and readability, reducing maintenance costs.
  • Accelerate Prototyping - Quickly implement ideas and iterate faster for validation.

Core Features

  • Unified Model Management

    Easily switch and combine different models by specifying the model provider via string.

  • Built-in OpenAI-Compatible Integration

    Built-in integration classes for OpenAI-Compatible APIs, enhancing model compatibility through explicit configuration.

  • Flexible Message Handling

    Supports chain-of-thought concatenation, streaming processing, and message formatting.

  • Powerful Tool Calling

    Built-in tool call detection, parameter parsing, and human-in-the-loop review functionality.

  • Efficient Agent Development

    Simplifies the agent creation process and provides more common middleware extensions.

  • Convenient State Graph Construction

    Provides two pre-built functions to facilitate the construction of sequential and parallel execution state graphs.

Quick Start

One of the main uses of this library is integrating models that provide OpenAI-Compatible APIs. Below is an example using qwen2.5-7b deployed via vLLM.

1. Install langchain-dev-utils

pip install -U "langchain-dev-utils[standard]"

Note: You must install the langchain-dev-utils[standard] version to use this feature.

2. Getting Started

from langchain.tools import tool
from langchain_core.messages import HumanMessage
from langchain_dev_utils.chat_models import register_model_provider, load_chat_model
from langchain_dev_utils.agents import create_agent

# Register model provider
register_model_provider("vllm", "openai-compatible", base_url="http://localhost:8000/v1")

@tool
def get_current_weather(location: str) -> str:
    """Get the current weather for a specified location."""
    return f"25 degrees, {location}"

# Dynamically load model using string
model = load_chat_model("vllm:qwen2.5-7b")
response = model.invoke("Hello")
print(response)

# Create agent
agent = create_agent("vllm:qwen2.5-7b", tools=[get_current_weather])
response = agent.invoke({"messages": [HumanMessage(content="What is the weather in New York today?")]})
print(response)

GitHub Repository

Visit the GitHub Repository to view the source code and issues.