TheBlewish/Automated-AI-Web-Researcher-Ollama

What it solves

This project addresses the limitation of standard LLM chatbots that cannot perform deep, systematic research. Instead of relying on a single prompt-and-response cycle, it automates the process of breaking down a complex query into multiple research areas, searching the web, and documenting findings to provide a verifiable and comprehensive answer.

How it works

The tool uses local LLMs via Ollama to execute a structured research loop:

1. Planning: It analyzes a user query and generates five prioritized research focus areas.
2. Execution: For each area, it formulates search queries, scrapes relevant web pages, and saves all content and source URLs to a text file.
3. Iteration: It uses findings from previous searches to generate new, refined focus areas, allowing the research to evolve.
4. Synthesis: Once the user terminates the research (or the LLM determines it has enough info), the LLM reviews the entire collected dataset to generate a final summary.
5. Interaction: The user can then enter a Q&A mode to ask specific questions about the gathered research.

Who it’s for

It is designed for users who want to conduct thorough, automated online research using local models to maintain privacy or control, and for those who need a documented trail of sources for their findings.

Highlights

• Automated Research Planning: Automatically breaks queries into prioritized focus areas.
• Self-Improving Search: Generates new research paths based on information discovered during the session.
• Documented Research Trail: Saves all scraped content and source links into a text file for verification.
• Local Execution: Runs entirely via Ollama, supporting models like Phi-3 with large context windows.