AdsCrawl for LangGraph

SETUP METHOD 01

Add a web-reading node

Define a StateSchema, add one AdsCrawl node and compile a runnable graph.

1. 1

   Install dependencies

   Run this in your application directory.

   ShellCopy

   npm install adscrawl @langchain/core @langchain/langgraph zod

2. 2

   Configure your API key

   Set the variables in your server environment or local shell. Replace the placeholders with your own keys; keep them out of browser bundles and source control.

   ShellCopy

   export ADSCRAWL_API_KEY="your_adscrawl_api_key"

3. 3

   Save langgraph-web.mjs

   The graph accepts a URL and adds Markdown to state. Empty page content raises an error instead of passing an empty result downstream.

   JavaScriptCopy

   import AdsCrawl from "adscrawl";
   import { END, START, StateGraph, StateSchema } from "@langchain/langgraph";
   import { z } from "zod";
   
   const browser = new AdsCrawl({ apiKey: process.env.ADSCRAWL_API_KEY });
   const PageState = new StateSchema({
     url: z.string().url(),
     markdown: z.string().default(""),
   });
   
   export const graph = new StateGraph(PageState)
     .addNode("read_page", async ({ url }) => {
       const markdown = await browser.markdown({
         url, waitUntil: "domcontentloaded",
       });
       if (!markdown.trim()) throw new Error("The page returned no text");
       return { markdown };
     })
     .addEdge(START, "read_page")
     .addEdge("read_page", END)
     .compile();
   
   const result = await graph.invoke({ url: "https://example.com" });
   console.log({ source: result.url, markdown: result.markdown });

4. 4

   Run and connect the next step

   Expect result.url to retain the input and result.markdown to contain Example Domain. Insert a processing node between read_page and END when you add extraction or summarization; configure a model only if that node needs one.

   ShellCopy

   node langgraph-web.mjs