Open Deep Research - AI-Powered Autonomous Research Workflow

Description

This workflow automates deep research by leveraging AI-driven search queries, web scraping, content analysis, and structured reporting. It enables autonomous research with iterative refinement , allowing users to collect, analyze, and summarize high-quality information efficiently.

How it works

1. 🔹 User Input

   • The user submits a research topic via a chat message.

2. 🧠 AI Query Generation

   • A Basic LLM generates up to four refined search queries to retrieve relevant information.

3. 🔎 SERPAPI Google Search

   • The workflow loops through each generated query and retrieves top search results using the SerpAPI API.

4. 📄 Jina AI Web Scraping

   • Extracts and summarizes webpage content from the URLs obtained via SerpAPI.

5. 📊 AI-Powered Content Evaluation

   • An AI Agent evaluates the relevance and credibility of the extracted content.

6. 🔁 Iterative Search Refinement

   • If the AI finds insufficient or low-quality information , it generates new search queries to improve results.

7. 📜 Final Report Generation

   • The AI compiles a structured markdown report , including sources with citations.

Set Up Instructions

🚀 Estimated setup time : ~10-15 minutes

• ✅ Required API Keys:

  • SerpAPI → For Google Search results
  • Jina AI → For text extraction
  • OpenRouter → For AI-driven query generation and summarization

• ⚙️ n8n Components Used:

  • AI Agents with memory buffering for iterative research
  • Loops to process multiple search queries efficiently
  • HTTP Requests for direct API interactions with SerpAPI and Jina AI

• 📝 Recommended Enhancements:

  • Add sticky notes in n8n to explain each step for new users
  • Implement Google Drive or Notion Integration to save reports automatically

🎯 Ideal for:
✔️ Researchers & Analysts - Automate background research
✔️ Journalists - Quickly gather reliable sources
✔️ Developers - Learn how to integrate multiple AI APIs into n8n
✔️ Students - Speed up literature reviews

🔗 Completely free and open-source! 🚀