Websearch on multiple links

Questions
1

Hi
I am currently working on a web search logic. How do we perform web search for a food product ,wherein the requirement is , the agent needs to go to the internet and refer 1st popular link say official site and get missing product information from the input data, and if that link doesn’t have sufficient information then should go to 2nd link and if both donot have information then should go to 3rd popular link? Request your guidance if anyone has done this before.y! The fastest way to find solutions is by using the :magnifying_glass_tilted_right: search function at the upper right.
If your question hasn’t been asked before, please follow the template below. Skip the questions that are not relevant to you. →

Describe the problem/error/question

What is the error message (if any)?

Please share your workflow

(Select the nodes on your canvas and use the keyboard shortcuts CMD+C/CTRL+C and CMD+V/CTRL+V to copy and paste the workflow.)

Share the output returned by the last node

Information on your n8n setup

  • n8n version:
  • Database (default: SQLite):
  • n8n EXECUTIONS_PROCESS setting (default: own, main):
  • Running n8n via (Docker, npm, n8n cloud, desktop app):
  • Operating system:
1 Like
2

Hi @Megha_Pai i was working on a similar flow like that, how i have approached this is by first setting up some predefined domains like i use SerpAPI to search across different sites, so i have researched a list of domains that sureshot write the content related to my use case and then i only google search with respect to those domains, so that my agent would get information, although in your case you can do that too but instead of searching for 1 just make it go through a loop where it will extract information from every site until it gets satisfied information, just create a list and a search queries and map those with domains so every query gets searched across, and so use SerpAPI node to get search results and then use firecrawl to extract pages one by one and feed it to text classifier agent and there 2 fields repeat and found, so if the extracted content is not satisfactory agent would output in the repeat branch and cycle will continue until it finds the correct response and outputs in found branch, this is very practical and production scale method, what matters here is your system prompt and instructions you give there and they should be very broad else this would solve your problem.