v0.1.0 — Initial Public Release

🌳 Pi-tree v0.1.0 — Initial Public Release

AI-assisted reading and research app with tree-structured conversations. Load your books, news feeds, papers, or YouTube videos, and an AI reads them with you — not as flat Q&A, but as branching conversations that help you cross from confusion into comprehension.

Local-first, bring your own key. Runs on your machine with cloud APIs (DeepSeek, Gemini, Claude) or fully offline with Ollama.

✨ Highlights

Core

• 🌳 Tree-structured conversations — branch on semantic shifts, navigate with TOC + chat
• 📚 Multi-source support — books (EPUB/PDF/MOBI), news feeds (RSS), academic papers, YouTube videos
• 👥 Multi-user with per-user isolated sessions, config, and glossaries
• 🔌 Plugin architecture — @pi-tree/plugin-sdk for building extensions with typed services
• 🤖 MCP client bridge — connect external MCP servers for web search, translation, etc.

Reading Experience

• 📖 Interactive reading sessions with context-aware AI companion
• 🔀 Multiple session modes per source (reading, analysis, Q&A, custom)
• 📝 AI-powered dictionary with inline lookup and glossary management
• 🧮 LaTeX math rendering via KaTeX
• 📊 Mermaid diagram support in conversations

Infrastructure

• 🐳 Docker support with ghcr.io/shuowu/pi-tree image
• 🖥️ Electron desktop app (macOS, Linux, Windows)
• 📖 VitePress documentation site
• 🔄 Auto-migration pipeline with Drizzle ORM
• 🧪 Unit + E2E test suites (Vitest + Playwright)
• 📋 Session profiles — declarative YAML mapping source types to skills/extensions

🐳 Docker Image

docker pull ghcr.io/shuowu/pi-tree:0.1.0

Also available as ghcr.io/shuowu/pi-tree:0.1 and ghcr.io/shuowu/pi-tree:latest.

See the self-hosting guide for configuration options.

📦 Getting Started

See the documentation for installation and setup instructions.