gemini-search

To find research papers beyond simple keyword matching, this AI-powered literature discovery decomposes topics into sub-problems and naming variants, returning a structured list of candidate papers with summaries and metadata.

Yes, you can search for open source code included with research papers. The discovery process identifies code-enabled work and returns a normalized output list containing the specific code URL for relevant candidate papers.

You get a structured list of research papers with arXiv ID and DOI by providing a broad research topic. The output normalizes title, authors, year, venue, arXiv ID, DOI, code link, and a one-sentence contribution summary.

Literature discovery works with the Gemini CLI if MCP is unavailable. The execution path tries the Claude Code MCP tool first and automatically falls back to the Gemini CLI to ensure continuous research search functionality.

The best way to discover survey papers and benchmarking starting points is using AI-powered search to surface diverse shortlists across multiple angles. It handles exploratory queries to return recent papers and surveys beyond traditional indexes.

To filter literature discovery results by year and venues, you parse arguments within your query. You can specify year ranges like 2022- and target specific venues to constrain the returned structured multi-attribute list of candidate papers.