619dev/guanlan

Guanlan is a modern RSS reader modeled after the now-discontinued Google Reader, built to feel quiet and clear while keeping you in control of your own data. It runs as a Cloudflare Worker and stores your subscriptions, articles, and read or starred status in a Cloudflare D1 database, so once deployed the whole thing lives on infrastructure you control rather than a third-party service. The interface uses a three-column layout: your list of subscribed feeds, the article list for whichever feed or filter you have selected, and a focused reading pane. You can filter by all articles, unread only, starred items, or a single feed, search across articles, switch the list density, and mark everything as read in one action. Adding a new subscription just takes an RSS or Atom feed address. The interface supports both Chinese and English, and offers light and dark themes that are remembered between visits. Access is protected by a password checked on the Worker's server side, with a signed, HttpOnly session cookie once you log in. It ships with a curated starter directory of 76 RSS feeds spanning Chinese and international tech media, developer communities, company engineering blogs, general news, and science publications, so a fresh deployment is not an empty inbox. This starter list is versioned, so redeploying with an updated catalog adds new feeds in small batches without overwriting feeds you already added yourself. Under the hood it is built with React 19, TypeScript, Tailwind CSS, and Vite, using a tool called Vinext to combine React Server Components with Worker-compatible builds, and Drizzle as the ORM layer over D1. Setup for local development means installing dependencies with npm, copying an example environment file, and setting your own access password and session secret before running the dev server. Deployment to Cloudflare Workers can be done by forking the repository and connecting it through Cloudflare's own dashboard, without ever committing secrets or database IDs into the repo, the database and its tables are created automatically on first deploy and first successful login. The project is released under the MIT license.