AI 驱动的智能书签管理系统

TMarks 是一款现代化的智能书签管理系统，旨在通过集成 AI 技术简化和提高书签管理的效率。它通过自动生成标签、多维筛选和标签页组管理等功能，为需要管理大量书签的用户（包括开发者和研究人员）提供一个高效的解决方案。

How It Works

TMarks 采用 React、TypeScript 和 Vite 构建前端，后端则基于 Cloudflare Workers 和 Pages Functions，利用 Cloudflare D1 (SQLite) 作为数据库，Cloudflare KV 进行缓存，并可选 Cloudflare R2 存储网页快照。其核心亮点是利用 OpenAI、Anthropic 等多种 AI 提供商自动为书签生成标签，并支持标签页组的智能收纳与恢复。这种 Serverless 架构和 AI 集成提供了可扩展、低运维成本的解决方案。

Quick Start & Requirements

• Local Development: Clone the repository, install dependencies (pnpm install), set up a local D1 database (wrangler d1 create tmarks-prod-db --local, pnpm db:migrate:local), and start the dev server (pnpm dev).
• Browser Extension: Navigate to the tab directory, install dependencies (pnpm install), and run in dev mode (pnpm dev).
• Deployment: Requires a Cloudflare and GitHub account. Deploy via Cloudflare Pages, connecting to a forked repository. Build command: pnpm build:deploy. Requires creating and binding D1, KV (recommended), and R2 (optional for snapshots) resources within Cloudflare. Sensitive environment variables (JWT secrets, encryption keys) must be configured securely in the Cloudflare dashboard.
• Prerequisites: Node.js, pnpm, Wrangler CLI, Cloudflare account, GitHub account.
• Links: Online Demo, Video Tutorial.

Highlighted Details

• AI-powered automatic tagging supports 8+ providers including OpenAI, Anthropic, DeepSeek, and Zhipu.
• Features include tab group management, public sharing pages, and a browser extension with offline support.
• Leverages Cloudflare's edge computing platform (Workers, Pages, D1, KV, R2) for a scalable, serverless architecture.
• Data security is addressed via JWT authentication, API key management, and data encryption.

Maintenance & Community

The project actively uses GitHub Issues for bug reporting and GitHub Discussions for feature suggestions. Specific contributor or sponsorship details are not highlighted in the README.

Licensing & Compatibility

The project is licensed under the MIT License, permitting commercial use and modification with attribution.

Limitations & Caveats

The snapshot feature is optional and requires R2 storage configuration; without it, this functionality is unavailable. The README does not explicitly detail other known limitations, alpha/beta status, or potential bugs.