Why LangChain.js? Chat Models and Prompt Templates

Everything LangChain does in Python, standardizing prompts, models, and chains behind one interface, has a first-class TypeScript port: LangChain.js. If you already know TypeScript's type system, this course reuses that strength directly, chains here are typed end to end.

Chat models

import { ChatOpenAI } from '@langchain/openai';

const model = new ChatOpenAI({ model: 'gpt-4o-mini', apiKey: 'YOUR_API_KEY' });
const response = await model.invoke('What is a large language model?');
console.log(response.content);

This needs a real API key and network access, so it's read-only here. model.invoke(...) is async, every LangChain.js call that talks to a model returns a Promise, matching how you'd already call any other network API in TypeScript.

Prompt templates

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

const prompt = ChatPromptTemplate.fromMessages([
  ['system', 'You are a helpful assistant that explains topics simply.'],
  ['human', 'Explain {topic} in two sentences.'],
]);

const messages = await prompt.invoke({ topic: 'recursion' });
console.log(messages);

{topic} is a placeholder, filled in from whatever object you pass to .invoke(), the same prompt object is reusable with { topic: 'closures' } or any other value, without rewriting the wording.

Why not just template strings yourself?

You could interpolate a plain TypeScript template string instead, and for a single, throwaway prompt, that's fine. ChatPromptTemplate starts paying off once you need to: reuse the same prompt in multiple places, compose it with a model and parser (next lesson), or swap the underlying model without touching the prompt at all.