What is this?
PaperRead is a personal Traditional Chinese research workspace for discovering papers, managing a candidate list, and planning reading paths. It's a fork of EasyRead that keeps dual-language reading, PDF layout, annotations and AI chat while adding research exploration and reading-planning tools on top.
PaperRead is a personal Traditional Chinese research workspace for discovering papers, managing a candidate list, and planning reading paths. It's a fork of EasyRead that keeps dual-language reading, PDF layout, annotations and AI chat while adding research exploration and reading-planning tools on top.Is this a desktop app or does it run in a browser?It's mainly an Electron desktop app with a bundled Python backend (installers for Windows, macOS and Linux), but there's also a Docker-based browser version that runs PaperRead as a web service with data persisted in a Docker volume.
It's mainly an Electron desktop app with a bundled Python backend (installers for Windows, macOS and Linux), but there's also a Docker-based browser version that runs PaperRead as a web service with data persisted in a Docker volume.How do I install and run PaperRead?Download the prebuilt installer for your OS from the Releases page (e.g. PaperRead-Setup-x.x.x.exe for Windows, the AppImage for Linux, or the arm64 DMG for macOS), use the provided Docker Compose setup for the browser version, or build from source with Python 3.10+ and Node.js 22+.
Download the prebuilt installer for your OS from the Releases page (e.g. PaperRead-Setup-x.x.x.exe for Windows, the AppImage for Linux, or the arm64 DMG for macOS), use the provided Docker Compose setup for the browser version, or build from source with Python 3.10+ and Node.js 22+.Do I need an AI API key to use it?AI-powered features like chat, summaries, smart annotation and mind maps need either a cloud API key (OpenAI, Anthropic, Google Gemini, DeepSeek, GLM, Kimi, OpenRouter, Groq, Cerebras, etc.) or a local model via Ollama/LM Studio, but basic search and the reading board keep working without AI.
AI-powered features like chat, summaries, smart annotation and mind maps need either a cloud API key (OpenAI, Anthropic, Google Gemini, DeepSeek, GLM, Kimi, OpenRouter, Groq, Cerebras, etc.) or a local model via Ollama/LM Studio, but basic search and the reading board keep working without AI.Any install pitfalls to watch out for?The Windows installer isn't code-signed yet, so Windows may flag it as an unverified publisher (functionality isn't affected); the macOS build only covers Apple Silicon, so Intel Macs need the Docker version or a from-source build; and Linux releases only ship an x64 AppImage.
The Windows installer isn't code-signed yet, so Windows may flag it as an unverified publisher (functionality isn't affected); the macOS build only covers Apple Silicon, so Intel Macs need the Docker version or a from-source build; and Linux releases only ship an x64 AppImage.Where does my paper library and notes get stored?The paper library and settings are stored locally on your machine for the desktop builds, or in a dedicated Docker volume for the browser version, and legacy EasyRead data is migrated automatically on first launch.
The paper library and settings are stored locally on your machine for the desktop builds, or in a dedicated Docker volume for the browser version, and legacy EasyRead data is migrated automatically on first launch.What's the difference between the candidate list, the reading board, and organized summaries?They're kept as separate states: the candidate list holds papers worth exploring later, the reading board tracks actual reading status (to-read/reading/to-organize/done), and organized summaries are saved permanently even after a paper leaves the candidate list.
They're kept as separate states: the candidate list holds papers worth exploring later, the reading board tracks actual reading status (to-read/reading/to-organize/done), and organized summaries are saved permanently even after a paper leaves the candidate list.Can it generate a reading path for a new research topic?Yes — starting from a topic or a specific paper, PaperRead orders the papers it actually retrieved into a sequence, labeling each one's role (foundational, method, representative, extension, frontier) and why it's recommended.
Yes — starting from a topic or a specific paper, PaperRead orders the papers it actually retrieved into a sequence, labeling each one's role (foundational, method, representative, extension, frontier) and why it's recommended.What is the Mermaid mind map feature?On the paper reading page you can generate a Traditional Chinese Mermaid mind map that visualizes the paper's research logic (question, method, findings, contributions, limitations), built only from text and data present in that specific paper.
On the paper reading page you can generate a Traditional Chinese Mermaid mind map that visualizes the paper's research logic (question, method, findings, contributions, limitations), built only from text and data present in that specific paper.What AI models can I use, cloud or local?You can connect cloud APIs like OpenAI, Anthropic, Google Gemini, DeepSeek, GLM, Kimi, OpenRouter, Groq and Cerebras, or run local models through Ollama or LM Studio, including custom OpenAI-compatible endpoints.
You can connect cloud APIs like OpenAI, Anthropic, Google Gemini, DeepSeek, GLM, Kimi, OpenRouter, Groq and Cerebras, or run local models through Ollama or LM Studio, including custom OpenAI-compatible endpoints.What are similar projects?easyread, qiaomu-clipper, get-it, cherry-studio, surf, opensquilla, kiss-translator, TradingAgents-CN
easyread, qiaomu-clipper, get-it, cherry-studio, surf, opensquilla, kiss-translator, TradingAgents-CNWhich topics is it in?mind-map, katex, chinese-traditional, bilingual, knowledge-management, note-taking, translation, academic-research, apple-silicon, ollama, electron, pdf
mind-map, katex, chinese-traditional, bilingual, knowledge-management, note-taking, translation, academic-research, apple-silicon, ollama, electron, pdf