Smart AI Content Research System Fully Automated Deep Research with n8n (free template)

---

---

AI Summary

Creating a Reusable Workflow with Google Search

Overview

• Utilize Google search for free in automations.
• Aim for 100 free searches daily instead of 100 monthly.

Steps to Create the Search Engine

1. Create Search Engine:
   • Go to the Programmable Search Engine (link in description).
   • Name it “AI Agents A2Z.” Configure to search the entire internet, excluding image searches.
2. API Configuration:
   • Enable programmatic access to get an API key.
   • Set up a project if needed, then copy the generated API key.

API Integration

• Required parameters:
  • key: API key.
  • cx: Search engine ID.
  • q: Search term.
• Useful optional parameters:
  • gl: Geolocation.
  • num: Number of search results (max 10).

Workflow Setup in Automation

1. Trigger: Start with a manual trigger; adjust later.
2. Edit Fields Node:
   • Set variables for API key, search engine ID, number of top results, context, and number of search terms.
3. LLM Chain:
   • Instruct LLM to generate search terms.
   • Example search terms: “Best documentaries 2024,” “Top documentary series released in 2024.”
4. Running Google Search:
   • Use HTTP request node with appropriate parameters for the Google search request.
   • Include site filters to avoid unwanted sites (e.g., exclude Reddit).
5. Filtering Results:
   • Filter to retain only HTML content.
   • Use custom code to clean up markdown output.
6. Summarization:
   • Use LLM to summarize aggregated content.
   • Include a step to assess relevance to the original context.
7. Final Output:
   • Aggregate results into a structured output for easy use.

Conclusion

• The workflow allows for deep research using Google search, leveraging a combination of LLMs and automation tools to create efficient search processes. Daily searches are limited to 100 for free, while exceeding this incurs costs.