Resources · n8n integration

AI Scraper n8n integrationAutomate AI Scraper with n8n.

Which page would be worth reading as clean rows instead of raw HTML? The AI Scraper n8n integration adds one node with 3 resources and 5 operations that send a URL or pasted markup to the Parsera API and return described fields, split across Extractor, Scraping Agent and Reusable Scraper.

Community node This node is not shipped with n8n: it is a community package (n8n-nodes-aiscraper) you install from Settings › Community nodes on a self-hosted instance. npm

Try the AI Scraper integration on n8nSee the 5 operations

Verified Trustpilot reviews · AI, automation & growth agency

Why automate

What does the AI Scraper n8n integration actually do?

AI Scraper is a node for n8n that calls the Parsera API. Instead of writing selectors, you describe the columns wanted and the service returns them as JSON items. The node covers 3 resources, Extractor, Scraping Agent and Reusable Scraper, for a total of 5 operations. There is no AI Scraper trigger node, so every workflow around it starts with something else.

The first pattern is a watch list. A Schedule Trigger fires on a fixed interval, a list of competitor pages runs through url-extractor.scrapeUrl with columns named price, availability and title, and each returned item lands in a spreadsheet. Because the extraction is described in plain words, a redesign on the target site does not break the run the way a CSS selector would.

The second pattern reuses HTML you already have. A workflow that fetches a page with the HTTP Request node, or receives markup inside a payload, can hand that text to url-extractor.parseHtml without a second fetch. One request out, one parse, no duplicate traffic to the source site.

The third pattern feeds a model. url-extractor.extractMarkdown turns a page into clean markdown text, which is far easier to summarize or embed than a wall of tags. Chain it with a summary step and post the result to Slack, or append it to a running brief in Google Docs.

When an endpoint of the Parsera API is not covered by the 5 operations, the HTTP Request node calls it directly and reuses the same credential through predefined authentication. That keeps one place to rotate the secret instead of two.

Two limits are worth knowing before building. There is no trigger, so scheduling or an inbound n8n Webhook does the waking up. And the node is a community node: n8n Cloud only accepts verified community nodes, so a self-hosted instance is the safe home for it. If a broader tour of the editor helps first, the n8n review and the n8n training cover the parts this page assumes.

Connect

How do you install and connect AI Scraper in n8n?

  1. 01

    Install the community package

    AI Scraper is not bundled with n8n. On a self-hosted instance, open Settings, go to Community nodes and install the npm package n8n-nodes-aiscraper, version 0.3.1. Once the install finishes, the node appears in the editor panel under the name AI Scraper, with its internal name n8n-nodes-aiscraper.aiScraper. Nothing is billed by n8n for a community node.

  2. 02

    Create the credential once

    Open the Credentials menu in n8n and create the AI Scraper credential there, following the Parsera user manual linked from the node documentation. A credential lives outside any single workflow: it is created one time and reused everywhere, so a later rotation touches one record rather than every node that calls the service. Save it before going back to the canvas.

  3. 03

    Pick the resource and the operation

    Drop the node into a workflow and choose the credential from its dropdown, which lists the ones already saved. Then pick the resource, Extractor, Scraping Agent or Reusable Scraper, and the operation under it. Fields marked required block execution when left empty, so a URL coming from an earlier node is written as an expression such as {{ $json.url }}.

Actions

What can the AI Scraper node do?

The AI Scraper node exposes 5 operations across 3 resources. For each one: the node as you configure it in n8n, the required fields, and our field notes.

Resources × operations matrix
ResourceInteract and ExtractExtract MarkdownParse HTMLRun Configured ScraperExtract URL
Extractor✓✓✓
Scraping Agent✓
Reusable Scraper✓

Extractor

3 operations
01

Extract from URL

url-extractor.scrapeUrl

What you see in n8n

Notes & use cases

Point the node at a page address, describe the columns wanted, and it returns structured data pulled from that page.

Key parameters

  • URL: required, the address of the webpage to extract data from, usually an expression such as {{ $json.url }} when the address comes from a previous node.
  • Columns Input Mode: fields defines each column in separate fields, json defines them with a JSON Schema, which suits complex schemas.
  • Mode: standard balances speed and accuracy, precision reaches data hidden inside HTML structures.
  • Proxy Country: routes the request through a proxy in the selected country to reach geo specific content.
Use cases
a nightly pass over a list of product pages, writing name, price and stock into Google Sheets.
02

Parse HTML

url-extractor.parseHtml

What you see in n8n

Notes & use cases

This one skips the fetch. Raw HTML or plain text already present in the workflow goes in, described fields come out.

Key parameters

  • Content: required, the raw HTML or text content to extract data from, typically the body returned by an earlier HTTP Request node.
  • Prompt (Optional): context and general instructions about the markup, useful when the same tags mean different things on different sections.
  • Schema (JSON): columns as a JSON object where each key is a field name and its value carries a type among any, string, integer, number, bool, list and object.
  • Mode: precision when the values sit deep in the structure.
Use cases
parsing an email newsletter body captured by Gmail into dated entries.
03

Extract markdown from URL

url-extractor.extractMarkdown

What you see in n8n

Notes & use cases

No columns to declare here. The operation converts a URL to clean markdown text, keeping the reading order of the page and dropping the markup noise.

Key parameters

  • URL: required, the address of the webpage to convert, often built from a link found in a previous item.
  • Proxy Country: the selected country routes the request through a proxy, which matters when the page serves different text per region.
Use cases
preparing documentation pages for a summary step or for indexing, where markdown is far cheaper to process than a full HTML document and keeps headings intact.

Scraping Agent

1 operation
04

Navigate interact and extract any data

scraping-agent.agentExtract

What you see in n8n

Notes & use cases

Some values only appear after a click or a scroll. This operation lets an AI agent navigate the page, interact with it and extract what the instruction asks for.

Key parameters

  • URL: required, the entry page the agent starts from.
  • Prompt: required, a description of what data to extract, in the spirit of extracting all product names and prices.
  • Define Output Schema: switched on, the returned data follows columns declared through Columns or Schema (JSON) instead of coming back loose.
Use cases
reading a catalogue that only reveals its later entries behind a load more button.

Reusable Scraper

1 operation
05

Run scraper configured on parsera org

reusable-scrapers.runScraper

What you see in n8n

Notes & use cases

When the extraction is already defined on parsera.org, n8n does not redefine it. This operation runs that saved scraper and returns its output.

Key parameters

  • Scraper Name or ID: required, picked from the list of existing scrapers, or set as an ID through an expression.
  • URL: optional, and left empty the scraper keeps its own default URL configuration.
  • Cookies: a JSON array of objects carrying name, value and domain, for pages that only answer to a session.
Use cases
one scraper maintained by the data team, called from several workflows without copying its settings.
Need help

Need help automating AI Scraper with n8n?

A person reads every message.

FAQ

AI Scraper and n8n, common questions

01Is the AI Scraper n8n integration free?
On the n8n side, yes. AI Scraper is a community node, which means it does not ship with n8n and costs nothing to install: the npm package n8n-nodes-aiscraper, version 0.3.1, is installed from Settings and then Community nodes on a self-hosted instance. n8n itself runs either self-hosted through Docker or npm, or on n8n Cloud, the hosted offer from n8n, and a workflow behaves the same way in both. One caveat on Cloud: only verified community nodes are accepted there. What the Parsera service itself costs is a separate matter and is not covered here.
02What credentials do you need for AI Scraper in n8n?
One credential, created once. In n8n, credentials live in their own Credentials menu rather than inside a workflow, so the record is created a single time and then offered in the dropdown of every AI Scraper node you add. That also means a rotation happens in one place. The exact fields to fill are described in the Parsera user manual that serves as documentation for this node and for its credential, so it is worth opening that page next to n8n while creating the record. Once saved, the credential is selected on the node and the resource and operation can be configured.
03What are the limits of the AI Scraper node in n8n?
The node exposes 5 operations across 3 resources, and that is the whole surface. Anything the Parsera API offers beyond those operations is reached with the HTTP Request node, which calls any endpoint of the API and reuses the same credential through predefined authentication, so there is no second secret to manage. Two more limits shape the design: there is no AI Scraper trigger node at all, and being a community node the package installs on a self-hosted instance while n8n Cloud accepts verified community nodes only. Plan the hosting before planning the workflow.
04Does AI Scraper react in real time in n8n?
No, because the integration has no trigger node. A workflow built around AI Scraper starts with something else: a Schedule Trigger running at a fixed interval, which fits watch lists and recurring extractions, an n8n Webhook node whose URL a third party service calls when it has something new, or the trigger of another tool already in the stack. The AI Scraper node then sits downstream and runs once per incoming item. In practice the schedule is the common choice, since the pages being read rarely announce their own changes.
05n8n or Make for AI Scraper?
It depends on four things. Hosting: n8n runs self-hosted through Docker or npm, or on n8n Cloud, while Make is hosted only. Data control: a self-hosted n8n keeps the scraped pages and the credential on your own infrastructure. Cost model: Make bills per operation, which adds up when a run extracts many rows, whereas a self-hosted n8n does not meter executions the same way. Visual logic: both editors are visual, with different habits around branching and iteration. For AI Scraper specifically, the community node points to self-hosted n8n, since it installs from Settings on that kind of instance.
Same platform

Other docs you might need on this stack.

Browse all integrations