中文文档

EasyPaper

Upload academic PDFs, get AI-powered analysis, chat about the content, and take sentence-level notes. Fully customizable API.

Features

• PDF Upload & Viewing — Drag-and-drop upload with built-in PDF viewer (zoom, page navigation, text selection).
• AI Analysis — Automatically extract summary, key contributions, methodology, and conclusions via MLLM (e.g. GPT-4o), with customizable prompts and batch progress tracking.
• Interactive Chat — Ask follow-up questions with full paper context. Supports multiple chat sessions per paper.
• Sentence-Level Notes — Select any text in the PDF and create notes linked to specific sentences, with tag management and Markdown support.
• Bookmarks — Mark important pages for quick navigation.
• Folder Organization — Organize papers in a hierarchical folder structure.
• Custom Prompts — Configure your own prompts for vision parsing, analysis, and chat (with Chinese/English presets).
• Theme Customization — Choose from 4 theme presets (dark-minimal, light-minimal, warm-light, warm-dark).
• Flexible AI Backend — Works with any OpenAI-compatible API (OpenAI, OpenRouter, etc.).
• Local Storage — All data stored locally in ~/.easypaper/, no external database needed.

Quick Start

Option 1: npx (no install)

npx @lvxintao/easypaper

Option 2: Global install

npm install -g @lvxintao/easypaper
easypaper

Option 3: From source

git clone https://github.com/lvxintao/EasyPaper.git
cd EasyPaper
npm install
npm run dev

Option 4: Update existing installation

npm i -g @lvxintao/easypaper@latest

Usage

# npm install
easypaper # Default port 3000, optional --port <number>

# From source
cd EasyPaper
npm run start

Open http://localhost:3000 to get started.

Configuration

Configure your AI provider in the app's Settings page, or via environment variables:

| Variable | Description | Default | |----------|-------------|---------| | AI_BASE_URL | API endpoint | https://api.openai.com/v1 | | AI_API_KEY | Your API key | — | | AI_MODEL | Chat model | gpt-4o | | AI_VISION_MODEL | Vision model for PDF parsing | gpt-4o |

• Environment variables: Can also be set in ~/.easypaper/.env, with lower priority than UI settings.

How to Use

1. Upload — Upload a PDF on the home page (drag-and-drop or click to browse)
2. Open — Enter the paper detail page with PDF viewer on the left and analysis panel on the right
3. Analyze — Get a structured analysis with page-referenced insights
4. Chat — Ask specific questions about the paper content
5. Take Notes — Select text in the PDF and create notes linked to specific sentences

CLI Options

easypaper [options]

  -p, --port <number>  Port to run on (default: 3000)
  -h, --help           Show help
  -v, --version        Show version

All data is stored locally in ~/.easypaper/.

Roadmap

• [x] Support sentence-level notes with PDF text selection
• [x] Support asking AI questions about specific sentences
• [ ] Support more file formats (currently only PDF)
• [ ] Support more PDF parsing methods (currently only MLLM)
• [ ] Support shared reading of papers
• [ ] Support asking AI questions about screenshots

Tech Stack

• Next.js 16 (App Router) + React 19 + TypeScript
• Tailwind CSS 4
• react-pdf for frontend PDF rendering
• mupdf for backend PDF page rendering

License

MIT License