Langsmith-sdk: Client SDK for LLM Debugging, Evaluation, and Monitoring

Introduction

The langsmith-sdk repository hosts the official Python and JavaScript SDKs for seamless interaction with the LangSmith platform. Developed and maintained by LangChain, the company behind the popular LangChain framework, LangSmith is designed to help teams debug, evaluate, and monitor their language models and intelligent agents. It integrates natively with both LangChain Python and LangChain JS open-source libraries, supporting any LLM application.

Installation

Getting started with the langsmith-sdk is straightforward for both Python and JavaScript environments.

Python

To install the Python SDK, use pip:

pip install -U langsmith

Then, set up your environment variables:

export LANGSMITH_TRACING=true
export LANGSMITH_API_KEY=ls_...
export LANGSMITH_WORKSPACE_ID=<your-workspace-id> # Required for org-scoped keys

JavaScript / TypeScript

To install the JavaScript/TypeScript SDK, use yarn or npm:

yarn add langsmith
# or npm install langsmith

And configure your environment:

export LANGSMITH_TRACING=true
export LANGSMITH_API_KEY=ls_...
export LANGSMITH_WORKSPACE_ID=<your-workspace-id> # Required for org-scoped keys

Examples

Once installed and configured, you can start tracing your LLM applications.

Python Example

import openai
from langsmith import traceable
from langsmith.wrappers import wrap_openai

client = wrap_openai(openai.Client())

client.chat.completions.create(
    messages=[{"role": "user", "content": "Hello, world"}],
    model="gpt-3.5-turbo"
)

JavaScript Example

import { OpenAI } from "openai";
import { traceable } from "langsmith/traceable";
import { wrapOpenAI } from "langsmith/wrappers";

const client = wrapOpenAI(new OpenAI());

await client.chat.completions.create({
  model: "gpt-3.5-turbo",
  messages: [{ content: "Hi there!", role: "user" }],
});

Why Use LangSmith?

LangSmith is an invaluable tool for anyone developing with Large Language Models. It provides a comprehensive platform for:

• Debugging: Quickly identify and resolve issues in your LLM applications.
• Evaluation: Systematically assess the performance and quality of your models and agents.
• Monitoring: Keep track of your LLM applications in production, ensuring optimal performance and detecting anomalies.

By offering these capabilities, LangSmith helps developers build more reliable, performant, and robust LLM-powered solutions.

Links

• GitHub Repository: langchain-ai/langsmith-sdk
• LangSmith Platform: smith.langchain.com
• LangSmith Documentation: docs.smith.langchain.com
• LangSmith Cookbook: langchain-ai/langsmith-cookbook
• LangChain Python: langchain-ai/langchain
• LangChain JS: langchain-ai/langchainjs