DeepResearch Assistant is a comprehensive prompt system designed to transform Large Language Models into rigorous research assistants. This system orchestrates a three-part workflow that mimics academic research processes to produce authoritative, evidence-based analyses on complex topics.
The first component focuses on comprehensive literature review, identifying high-quality academic sources, evaluating methodological rigor, and ensuring diverse scholarly perspectives.
The second component extracts, analyzes, and integrates complex information from multiple sources, organizing evidence by theme and strength while critically evaluating quality and limitations.
The final component crafts comprehensive analyses with precise research questions, structured logical frameworks, and evidence-based conclusions while maintaining appropriate academic rigor and citation practices.
- Copy the entire prompt system into your preferred LLM interface as a system prompt
- Modify the prompt as needed for your specific research needs
- Submit your research query as a user prompt
- The system will guide the LLM to produce a comprehensive research analysis following academic standards
- Prioritizes peer-reviewed and authoritative sources
- Evaluates methodological quality and evidence strength
- Identifies areas of scholarly consensus and disagreement
- Maintains appropriate epistemic humility
- Produces structured research outputs with proper citations
- Distinguishes between facts, consensus views, and interpretations
The system produces research reports with the following structure:
- Executive summary
- Background and context
- Methodology
- Evidence analysis by theme
- Synthesis and implications
- Future research directions
- References
- Effectiveness depends on the capabilities of the underlying LLM
- Requires internet access for research capabilities
- Limited by the LLM's knowledge cutoff date for recent developments if no internet access available
Contributions to improve the prompt system are welcome. Please feel free to submit issues or pull requests.