PaperSpark provides an AI-driven workbench for researchers and advanced learners, streamlining academic workflows from literature management and reading to knowledge synthesis and paper writing. It offers a unified solution for collecting, understanding, and producing academic content, aiming to enhance productivity and insight generation.

How It Works

The system employs a multi-agent AI architecture within a Next.js/React frontend. It facilitates a closed-loop workflow: literature import (local, Zotero) into a knowledge base, immersive dual-column reading with annotation, automatic knowledge graph construction, AI-assisted writing via a rich text editor with formula support, and multi-format export (Markdown, TeX, LaTeX Zip). Optional Python engines (local Surya, cloud Modal/MinerU) enhance OCR, layout analysis, and code execution capabilities.

Quick Start & Requirements

Requires Node.js 18+ and pnpm 8+. Basic functionality is available without Python. Advanced features like OCR, layout analysis, and code execution necessitate Python 3.10+ (with optional CUDA 12.x for GPU acceleration via Surya) or cloud services (Modal, MinerU). Documentation is available at https://docs.paper.062679.xyz. Docker deployment is also supported.

Highlighted Details

• AI Academic Assistant: Multi-agent system for retrieval, rewriting, error correction, translation, summarization, and code execution.
• Immersive Reading: Dual-column view, focus extraction, chapter mind maps, annotation, and paper reference link navigation.
• Knowledge Graph: Automatic generation and visualization of paper, author, and concept relationships.
• Export Formats: Supports Markdown, TeX, and a compilable LaTeX Zip package.

Maintenance & Community

The project is in its early stages ("建设初期"). Community interaction is encouraged via QQ group 1082678889. Contributions follow Conventional Commits.

Licensing & Compatibility

Licensed under CC BY-NC 4.0. Commercial use, paid distribution, or integration into commercial products is strictly prohibited without explicit authorization.

Limitations & Caveats

Documentation and environment setup steps are incomplete. Advanced parsing and execution capabilities are optional and depend on external Python environments or cloud services. Commercial adoption is blocked by the non-commercial license.