A webpage in. Useful data out.

Download a complete workflow

Save a public webpage as a Markdown file in Google Drive. The workflow runs manually, checks the extraction, and stops for errors or warnings.

Download the n8n workflow JSON · Setup guide and source

1. In n8n, create a workflow and choose Import from File in its menu.
2. Import the downloaded JSON and enter your URL in Website URL.
3. Connect Header Auth on Extract with FetchRelay using the instructions below.
4. Connect Google Drive OAuth and select the folder you want to save to.
5. Execute manually. Each successful run uses one FetchRelay credit and creates a new file.

Start with Example Domain to check the setup, then try your own public page. Inspect any warnings before enabling a recurring workflow.

Before you start

You need an n8n workspace and a FetchRelay API key from Account. This is a direct HTTP integration recipe, not an official n8n marketplace node or partnership. n8n hosting and subscriptions are separate from FetchRelay. Successful page extraction uses an account credit.

1. Add an HTTP Request node

Start a workflow with a manual trigger and add HTTP Request. Set Method to POST and URL to https://fetchrelay.com/api/scrape. Keep the first workflow inactive while testing so a schedule does not repeatedly consume credits.

2. Store the key as a credential

Choose generic authentication with Header Auth. Set the credential’s header name to Authorization and its value to Bearer followed by a space and your FetchRelay API key. Select that credential on the node. Do not paste a real key into shared workflow JSON, exported examples, screenshots, or the request URL.

3. Send a JSON body

Enable the request body, select JSON, and enter:

{
  "url": "https://example.com",
  "formats": [
    "markdown",
    "links"
  ],
  "mode": "http",
  "timeout": 25000
}

Set the HTTP node’s timeout longer than 25 seconds; 40 seconds is a reasonable starting value for this example. Keep its normal HTTP error handling enabled.

4. Execute and check the result

Run the node once. Expect success: true, the Example Domain heading in data.markdown, and a data.links array. Check diagnostics.warnings before accepting the extraction. Use an IF node to separate success from error handling and to route qualified results for review.

The expressions below read the current node input in later workflow steps:

{{ $json.data.markdown }}
{{ $json.data.metadata.sourceURL }}
{{ $json.diagnostics.warnings }}

5. Connect your next step

Map the Markdown to your own document store, search index, or reporting step. For product fields, use explicit CSS selectors and validate the title, price, currency, and availability before writing a row. Begin with the test catalog example.

For a crawl, save the returned job ID and poll the status endpoint instead of repeating the create request. Collection jobs can finish with partial results. The crawl documentation explains job status and exports.

When something fails

Check the credential for authentication errors. Slow down on rate limits. Do not retry unavailable or bot-protected pages indefinitely. A client timeout can happen after the server has completed a request, so a retry may create another billable extraction. Browser mode can run JavaScript; it cannot promise access through login walls or CAPTCHA.

Reference: n8n HTTP Request documentation. This guide uses the same API contract as the FetchRelay OpenAPI reference.