How can I extract all product records from order confirmation emails using n8n's Information Extractor?

Below is one approach that has worked for others doing similar tasks with n8n:

1. Pre-Process the HTML with a Parser

Instead of relying on the Information Extractor to pull out multiple products automatically, you can extract the table rows first. For example, you could add a node (like the Cheerio node or a Code node) to parse the HTML and select all rows in the products table. This way, each product row becomes its own item in the workflow.

2. Use a Loop Over Items Node

Once you have separated the table rows, you can then use the Loop Over Items node to iterate over each row. In each iteration, call the Information Extractor (or your OpenAI-based extraction) so that each product is processed individually.

3. Adjust the Prompt in the Information Extractor

If you wish to continue processing the whole email as is, you might need to modify the prompt provided to your OpenAI account. Make sure it instructs the extractor to return a list or an array (for example, a JSON list of products) rather than only one product. Explicitly stating “return all products” or “extract all table rows” can sometimes yield better results.

Summary

• Pre-Process the Email: Use a parser to extract each table row from the order confirmation.
• Loop Through Items: Use the Loop Over Items node to handle each product separately.
• Refine the Prompt: Update your extraction instruction so that OpenAI returns multiple items rather than stopping at the first match.

This approach leverages n8n’s strength in splitting and looping through data along with your extractor node to ultimately get a structured output for every product listed. Happy automating!