How to beginner · 3 min read

How to trace LangChain calls with LangSmith

Q: How to trace LangChain calls with LangSmith

To trace LangChain calls with LangSmith, set the environment variables LANGCHAIN_TRACING_V2 to true, LANGCHAIN_API_KEY to your LangSmith API key, and LANGCHAIN_PROJECT to your project name. This enables automatic tracing of all LangChain calls without code changes.

Quick answer

To trace LangChain calls with LangSmith, set the environment variables LANGCHAIN_TRACING_V2 to true, LANGCHAIN_API_KEY to your LangSmith API key, and LANGCHAIN_PROJECT to your project name. This enables automatic tracing of all LangChain calls without code changes.

PREREQUISITES

Python 3.8+
LangChain v0.2+ installed
LangSmith API key
pip install langsmith

Setup

Install the langsmith package and set environment variables to enable tracing.

Install LangSmith SDK: pip install langsmith
Set environment variables in your shell or .env file:

export LANGCHAIN_TRACING_V2=true
export LANGCHAIN_API_KEY=your_langsmith_api_key
export LANGCHAIN_PROJECT=my-project

bash

pip install langsmith

Step by step

Here is a complete example showing how to enable LangChain tracing with LangSmith and run a simple LangChain prompt.

python

import os
from langchain_openai import ChatOpenAI
from langchain_core.prompts import ChatPromptTemplate
from langchain.chains import LLMChain

# Environment variables must be set before running this script:
# LANGCHAIN_TRACING_V2=true
# LANGCHAIN_API_KEY=your_langsmith_api_key
# LANGCHAIN_PROJECT=my-project

# Initialize LangChain LLM
llm = ChatOpenAI(model="gpt-4o-mini", temperature=0)

# Create a simple prompt template
prompt = ChatPromptTemplate.from_template("Say hello to {name}!")

# Create a chain
chain = LLMChain(llm=llm, prompt=prompt)

# Run the chain
result = chain.invoke({"name": "LangSmith"})
print("Output:", result)

# All calls are automatically traced to LangSmith dashboard

output

Output: Say hello to LangSmith!

Common variations

You can also use LangSmith tracing with asynchronous LangChain calls or different LLM models. The tracing works automatically once environment variables are set.

Async example: use await chain.invoke_async(...) in an async function.
Use other LLMs like ChatAnthropic or ChatOpenAI with the same tracing setup.
Trace non-LangChain code by using the LangSmith Client and @traceable decorator.

python

import asyncio
from langchain_openai import ChatOpenAI
from langchain_core.prompts import ChatPromptTemplate
from langchain.chains import LLMChain

async def main():
    llm = ChatOpenAI(model="gpt-4o-mini", temperature=0)
    prompt = ChatPromptTemplate.from_template("Say hello to {name}!")
    chain = LLMChain(llm=llm, prompt=prompt)
    result = await chain.invoke_async({"name": "Async LangSmith"})
    print("Async output:", result)

asyncio.run(main())

output

Async output: Say hello to Async LangSmith!

Troubleshooting

If you do not see traces in the LangSmith dashboard, verify that LANGCHAIN_TRACING_V2 is set to true and your API key is correct.
Ensure you are using LangChain v0.2 or newer, as older versions do not support LangSmith tracing.
Restart your Python environment after setting environment variables.

✅

Key Takeaways

Set LANGCHAIN_TRACING_V2=true and LANGCHAIN_API_KEY to enable automatic LangChain tracing with LangSmith.
No code changes are needed; tracing works by environment configuration and LangChain v0.2+ integration.
Use langsmith.Client and @traceable decorator for manual tracing outside LangChain.
Async and different LLM models are supported with the same tracing setup.
Verify environment variables and LangChain version if traces do not appear in LangSmith.

Verified 2026-04 · gpt-4o-mini

Verify ↗