PaperFlow AI

A persistent AI research companion beside every paper.

PaperFlow is a local-first research workspace for reading, annotating, organizing, and discussing academic papers. Open arXiv and direct PDF links in an integrated reader while preserving the original source URL. Highlight, underline, strike through, draw, add comments and area notes, search documents, and run local OCR for scanned pages. Organize papers with nested collections, tags, favorites, and recently read views. Notes, annotations, conversations, and reading progress are stored locally. AI features are optional. Users can connect an existing Codex subscription through the local Native Host or configure a direct OpenAI-compatible API connection without installing the host. Content is sent only when the user explicitly invokes an AI action. Optional encrypted Google Drive synchronization keeps library records, notes, annotations, conversations, and selected offline PDFs available across devices. Data is encrypted on the device before upload. PaperFlow contains no advertising, analytics, tracking, or telemetry.