Research2026v0.1.0

Agentic Literature RAG

An evidence-grounded CLI for discovering, reading, and synthesizing academic literature.

Overview

Agentic Literature RAG is a cross-platform research workflow that discovers papers through arXiv, DOI lists, and local PDFs; enriches scholarly metadata; builds persistent local vector indexes; and produces structured literature reviews through OpenAI-compatible models. Its outputs include page-level citations, evidence matrices, citation audits, bibliographies, and ranked thesis proposals.

Current notes

  • Tiered paper discovery, relevance triage, and persistent research workspaces
  • Local CPU embeddings with safe, automatically invalidated FAISS indexes
  • Evidence-grounded synthesis with page-level citations and critic review
  • Markdown, JSON, BibTeX, paper manifest, and thesis proposal exports