Fetch Papers

Overview

Fetches the README of a GitHub list of research papers, keeps the links that point at PDFs, and downloads each PDF into the run workspace.

This tutorial does not ship as an importable template. Build it manually by following the steps below. NodeTool has no HTTP request or download node, so both network steps are Code nodes that call the sandbox fetch.

Demo

Workflow Steps

  1. Get README (nodetool.code.Code) - Calls fetch(url) on the raw README URL and reads text(). The same body pulls the link targets that end in .pdf and returns them as a list. Use the @nodetool-ai/sandbox-html pack’s extractLinks instead when the source is an HTML page.
  2. For Each (nodetool.control.ForEach) - Emits one URL at a time from input_list.
  3. Download (nodetool.code.Code) - Fetches each URL and saves it with workspace.writeBytes(name, await res.bytes()).
  4. Collect (nodetool.control.Collect) - Gathers the saved file names into a list for a Preview node.

fetch refuses loopback, link-local, and private addresses. See the JavaScript Sandbox reference for its limits.

Tags

automation

Workflow Diagram

graph TD readme["Code (get README, keep PDF links)"] each["ForEach"] download["Code (download)"] collect["Collect"] preview["Preview"] readme --> each each -->|output| download download --> collect collect --> preview
Edit this page on GitHub