Question 1

What is Recommendica and how does it work?

Accepted Answer

Recommendica is a research recommendation engine built on retrieval-augmented generation (RAG). You ask a research question in plain language; it expands your query across multiple representations, searches indexed academic papers and live arXiv literature, grades each paper for direct relevance, and writes a synthesized answer strictly grounded in the relevant papers.

Question 2

How does Recommendica prevent AI hallucinations?

Accepted Answer

Unlike standard conversational LLMs, Recommendica uses relevance grading, multi-chunk grounding, and claim-level verification. Every claim generated is scored against its source papers with an aggregate faithfulness score, and the engine stays silent or refines queries rather than guessing from unrelated documents.

Question 3

What paper repositories does Recommendica search?

Accepted Answer

Recommendica searches curated collections of scientific and academic literature across Artificial Intelligence, Computer Science, Climate Science, Biomedicine, Physics, and Mathematics, supplemented by real-time arXiv paper retrieval fallback.

Question 4

Is Recommendica free to use for researchers and students?

Accepted Answer

Yes, Recommendica is completely free to use with no account registration or paywall required.