Need help creating custom Flow

Hey everyone :waving_hand:,

I’m working on an automation with n8n + Flaskband I’d really appreciate some guidance. The idea is:

1. A user sends a request in plain text (something like “I need these items”).

2. An AI Agent in n8n extracts the relevant keywords from the request.

3. n8n has access to a local CSV file with a list of products and two price columns from different sources (e.g. Source A and Source B).

4. The workflow should match the extracted items with rows in the CSV (even partial matches are okay).

5. Then it should build a structured JSON like this:{ “items”: \[ { “name”: “product1”, “price\_source\_a”: 120, “price\_source\_b”: 115, “quantity”: 1 }, { “name”: “product2”, “price\_source\_a”: 350, “price\_source\_b”: 340, “quantity”: 2 } \], “total\_source\_a”: 470, “total\_source\_b”: 455 }

6. Finally, that JSON should be sent to my local Flask API as a POST request.

To build your workflow @Tentordev

  1. Use an AI node to get keywords from the user’s text.
  2. Load your product list from a CSV file.
  3. Use a function node to match keywords with products in the CSV.
  4. Create a structured JSON output with item details and totals.
  5. Use an HTTP node to send the JSON to your Flask API.

Break it down step by step, and you’ll get it working
Good luck

2 Likes

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.