SearchAgent

Archived learning project. Built to understand the "Search" layer of Search + RAG pipelines.

What it taught

• How search engines return structured results (title, URL, snippet)
• How to pass search results as context into an LLM prompt
• Difference between raw HTML and structured search results

Components

• search.py — DuckDuckGo search via ddgs, returns {title, url, snippet}
• llm.py — passes search results as LLM context
• templates/ + static/ — minimal web UI

Run

pip install ddgs
python search.py

Note

This project is archived. The requirements.txt lists packages (fastapi, openai, tavily) that were explored but are not used in the current code. Only ddgs is required.