curso-langchain/slides/pages/models/15.md

layout

layout
default

Ejemplo: Agente con Tavily Search

import { createAgent } from "langchain";
import { TavilySearchResults } from "@langchain/community/tools/tavily_search";

const searchTool = new TavilySearchResults({ apiKey: process.env.TAVILY_API_KEY, maxResults: 5 });
const agent = createAgent({
  model: "openai:gpt-4.1-mini",
  tools: [searchTool],
});

const result = await agent.invoke({
  messages: [{ role: "user", content: "¿Cuales son las ultimas noticias sobre IA?" }],
});

Tavily - Optimizado para LLMs: extrae contenido relevante, filtra anuncios/ruido, API key gratuita limitada

Docs: langchain.com/integrations/tools