Building a Simple LLM Application with LangChain

Overview

Objective: Create a simple application to translate text from English into another language using LangChain.
Components:
- Language Models
- Prompt Templates
- Debugging/Tracing with LangSmith

Setup

Installation

Install LangChain using package managers:
- npm
- yarn
- pnpm
Command: yarn add langchain @langchain/core
Installation Guide

LangSmith Setup

For complex applications with multiple LLM calls, use LangSmith for tracing.
Set environment variables for logging: export LANGSMITH_TRACING="true" export LANGSMITH_API_KEY="..." # Optional: Reduce tracing latency export LANGCHAIN_CALLBACKS_BACKGROUND=true

Using Language Models

LangChain supports various language models:
- Groq, OpenAI, Anthropic, FireworksAI, MistralAI, VertexAI

Example with Groq Model

Installation

Add dependency: yarn add @langchain/groq

Environment Variables

Set API key: GROQ_API_KEY=your-api-key

Model Instantiation

import { ChatGroq } from "@langchain/groq";
const model = new ChatGroq({  model: "llama-3.3-70b-versatile",  temperature: 0});

Using the Model

Models are Runnables.
Use .invoke method with a list of messages to interact with the model.

import { HumanMessage, SystemMessage } from "@langchain/core/messages";
const messages = [
  new SystemMessage("Translate the following from English into Italian"),
  new HumanMessage("hi!"),
];
await model.invoke(messages);

View token usage and other details via LangSmith if enabled.

Streaming

Stream tokens from a chat model using async and streaming modes:

const stream = await model.stream(messages);
const chunks = [];
for await (const chunk of stream) {
  chunks.push(chunk);
  console.log(`${chunk.content}|`);
}

Prompt Templates

Construct messages from user input and application logic.
Transform raw input into a prompt using Prompt Templates.

Creating a Prompt Template

Use variables for language and text.

import { ChatPromptTemplate } from "@langchain/core/prompts";

Define the system message:

const systemTemplate = "Translate the following from English into {language}";

Create the PromptTemplate:

const promptTemplate = ChatPromptTemplate.fromMessages([
  ["system", systemTemplate],
  ["user", "{text}"],
]);

Invoke the prompt template:

const promptValue = await promptTemplate.invoke({  language: "italian",  text: "hi!",});
promptValue;

Access messages:

promptValue.toChatMessages();

Invoke the model:

const response = await model.invoke(promptValue);
console.log(`${response.content}`);

Conclusion

Gained skills in working with language models, creating prompt templates, and using LangSmith for observability.
For further knowledge, explore Conceptual Guides and How-to Guides.

Creating a LLM Translation App with LangChain

Building a Simple LLM Application with LangChain

Overview

Setup

Installation

LangSmith Setup

Using Language Models

Example with Groq Model

Installation

Environment Variables

Model Instantiation

Using the Model

Streaming

Prompt Templates

Creating a Prompt Template

Conclusion