Researching complex topics is often tiresome. Sifting through countless search results, verifying sources, and synthesizing key insights takes significant time and effort. Deep Research simplifies this process by using AI to automate and enhance research workflows, delivering comprehensive, well-structured reports in minutes.

How Deep Research Works

Deep Research is built on top of a great open source project, which leverages powerful AI-driven techniques to refine research results iteratively. At its core, it combines:

• SERP (Search Engine Results Pages) exploration – It formulates multiple internet search queries to explore different angles of a topic.
• Firecrawl – A web crawler that retrieves relevant, high-quality content from various sources.
• Iterative AI-driven synthesis – AI models analyze and refine extracted information, improving the report step by step.

The tool doesn't just return a list of links—it curates, extracts, and organizes the most relevant insights into a coherent report.

Step by Step Process

1. User submits a research query – Any topic you need to explore.
2. AI generates clarifying questions – Refining the research scope.
3. The system crafts multiple search queries – Covering diverse perspectives.
4. Firecrawl retrieves relevant sources – Gathering high-quality content.
5. AI extracts key insights – Filtering and summarizing the most valuable findings.
6. Iterative refinement – AI polishes and structures the report for clarity.

The result is a detailed, well-organized Markdown report with proper citations and source tracking.

Choose Your AI Model

Deep Research offers multiple AI models to balance cost and performance:

Model	Cost (approx.)
GPT-4o	~$1 per run
GPT-4o mini	~$0.20 per run
O3 mini	~$0.40 per run

You can select the model that fits your needs and budget before launching a research session.

Where can I try it?

You can find Deep Research in our new "tools" section in the model dropdown

Try Deep Research Today 🚀