Skip to main content
The Follow links block is a fundamental navigation tool that tells the AI agent to visit a URL or a list of URLs. Finding links with a block like Find Links is not enough; you must use this block immediately after to explicitly instruct the agent to go to those links. This block acts as the bridge between finding a path and actually walking down it.

​
Purpose

Use the Follow links on a page block to:
  • Navigate to a list of URLs found by the Find Links block.
  • Open a URL that has been extracted as a variable by an Extract Data block.
  • Systematically visit a series of pages to perform an action or extract data from each one.

​
Configuration

Configuration is centered around telling the agent which links from the incoming list it should visit. This is a dropdown menu that controls the navigation behavior.
  • Follow the First link: The agent will only navigate to the very first URL from the list provided by the previous step.
  • Follow each link: The agent will iterate through the entire list of URLs, executing all subsequent blocks for each link. This is the most common option for scraping multiple pages.
  • Follow links that have not been seen before: A powerful option for monitoring workflows. The agent will keep a memory of links it has visited in previous runs and only follow links it has never seen before, preventing duplicate processing.
  • url_strings: Allows you to manually provide a URL or, more commonly, use a variable {{variable_name}} that contains a URL.
Follow Links block UI

The Follow links block configuration panel.

​
Common Workflow Patterns

This block is almost always used in one of two patterns. This is the standard way to scrape a list of items.
  1. Find Links: You first use Find Links to identify all the relevant <a> tag links on a page (e.g., all product links from a category page).
  2. Follow links: You immediately add this block next. By selecting Follow each link, you tell the agent to visit every single URL that was just found.
  3. Extract Data: After this, you add an Extract Data block to scrape information from each of the detail pages the agent visits.

​
Pattern 2: After Extract Data

This pattern is used when the URL you want to visit is not a direct <a> tag but is extracted as text from the page.
  1. Extract Data: Your first extraction step scrapes a piece of data and saves it to a variable with the specific key URL.
  2. Follow links: You add this block next. The agent will recognize the URL key from the previous step and navigate to the link contained in its value.

​
Key Considerations

  • Mandatory Preceding Block: This block is not a starting point. It must follow a block that provides it with a URL or a list of URLs, typically Find Links or Extract Data.
  • Iteration: When Follow each link is selected, it creates a loop. All subsequent blocks will be executed for every single link in the list.
  • De-duplication: The Follow links that have not been seen before option is extremely useful for building monitoring automations that run on a schedule and should only process new content.
The Follow links block is the essential engine for moving your agent from a list of possibilities to the actual pages you want to work on.