gemini-search

Broad literature discovery decomposes a research topic into multiple angles, aliases, and sub-problems to find relevant papers beyond simple keyword matching. This approach surfaces surveys, recent papers, and related variants for your query.

Yes, literature discovery can optionally prioritize papers with code. The search returns a structured response that includes a specific code URL field for each paper, allowing you to easily identify research with available open source implementations.

Gemini-powered search returns a structured response containing the title, authors, year, venue, arXiv ID, DOI, code URL, and a concise one-sentence summary for each paper. This normalizes outputs for easy comparison.

You can use either a Gemini MCP tool call or a Gemini CLI fallback to run broad literature searches. The MCP integration is preferred for tool use, but the CLI provides resilience with constraints like MAX_RESULTS and MIN_YEAR.

To collect a high-coverage starting set for an early-stage literature review, use broad literature discovery to search across multiple angles and venues. You can specify constraints like MAX_RESULTS and MIN_YEAR to filter the returned papers.

The best way to find papers across multiple venues and sub-problems is using Gemini-powered broad discovery. It decomposes your thesis topic into related variants and aliases, returning structured fields like arXiv ID and DOI for high coverage.