Autonomous multi-source research with configurable providers.
Deep Research is an autonomous research capability that breaks complex questions into parallel search tasks, scours multiple web sources, and synthesizes findings into a comprehensive report with citations. The AI launches a dedicated research agent that searches across multiple providers and synthesizes results.
When to use Deep Research
Deep Research is ideal for questions that require gathering and synthesizing information from many sources:
Literature reviews -- "What turbulence models are best for separated flow over a backward-facing step?"
Solver comparisons -- "Compare OpenFOAM's interFoam with the VOF implementation in Fluent for dam-break simulations."
Methodology questions -- "What are the recommended best practices for meshing automotive underbody aerodynamics?"
Standards and regulations -- "What CFD validation requirements does ASME V&V 20 specify?"
Deep Research vs. Web Search
Use Web Search for quick factual lookups. Use Deep Research when you need a thorough, multi-source investigation with synthesis and citations.
How it works
Submit your research question
Ask a question that benefits from in-depth investigation. You can trigger Deep Research explicitly by saying "research this in depth" or the AI may suggest it for complex questions.
Clarifying questions (optional)
The research agent may ask one or two clarifying questions before proceeding -- for example, which solver version you are using or what Reynolds number range you care about. This helps focus the search.
Parallel research tasks
The question is decomposed into sub-queries that run in parallel. Each sub-query searches different sources and angles of the problem.
Synthesis and report
Results from all sub-queries are combined into a structured report with headings, key findings, and inline citations linking back to the original sources.
Research progress
Deep Research runs in the background. When you submit a research request, the tool returns immediately and the client polls for results. While research is running, you see inline progress updates showing:
The sub-queries being explored
Sources being accessed
Intermediate findings as they emerge
You can continue chatting in the same conversation while research runs. The final report is posted when synthesis is complete.
Output format
The research report is delivered as a structured Markdown artifact in the chat. It typically includes:
Executive summary -- Key takeaways in 2-3 sentences
Detailed findings -- Organized by sub-topic with explanations
Citations -- Numbered references with links to original sources
Recommendations -- Actionable suggestions specific to your simulation context
Availability
Deep Research is available on all paid plans. Free-tier users can access it with sufficient credits. The feature requires an active internet connection since it queries live web sources.