Thirsty for Knowledge

Problem

Most blogs are one-directional: an author writes, readers consume, and there's no real channel for readers to shape what comes next. Thirsty for Knowledge was built to close that gap for research-driven, long-form writing.

Approach

Admins publish long-form articles through a rich text editor, and readers engage through likes, threaded comments, and shares. A separate topic request board lets readers pitch ideas and upvote or downvote existing ones, giving the author a ranked signal for what to write next. Google OAuth via BetterAuth handles authentication with role-based access separating admin and reader permissions.

Key Decisions

• Denormalized author name and image directly onto comment and topic request documents at creation time, after discovering BetterAuth's user collection isn't a registered Mongoose model and can't be .populate()d.
• Wrote a small custom sanitization middleware after express-mongo-sanitize turned out to be incompatible with Express 5's read-only req.query, avoiding a third-party dependency for a problem solvable in a few lines.
• Added a per-post language field with a client-side <html lang> switcher instead of pulling in a full i18n framework, since the site serves genuinely bilingual (Bengali/English) content post-by-post rather than a fully translated UI.
• Configured cross-domain cookies (sameSite: none, secure: true) to fix OAuth session handling once the frontend and backend moved to separate production domains.

Outcome

A fully deployed, production-ready blog platform with content moderation tools, a voting-driven topic pipeline, and SEO fundamentals (structured data, sitemap, dynamic per-post metadata) in place from day one.

What's Next

• Real-time comment and vote updates with Socket.io.
• AI-generated summaries for longer research articles.
• An admin analytics dashboard tracking post performance and engagement trends.