Graphlexity is an ai internet search engine.

It uses google serper to search internet and scrapes the top link(s) using crawl4ai.

Used technologies

• Crew ai for creating agents, tasks
• Langchain to create custom tools and general llm calls
• Google serper langchain tool to search internet. Which is also used as citations.

My initial idea was to scrape the web page in markdown format and pass the content to llm and get a summary, to make the citations as accurate as possible but because the scraped content is a lot and it consumes lots of openai credit, I'm not adding this feature.

The google serper web page snippet is passed to the llm for now.