Building an Automated Website Posting System
What We Did
Created an end-to-end posting system that lets our AI assistant generate blog posts about projects we build together. The system includes:
- Post template with metadata (title, date, tag, excerpt, slug) - Markdown-to-HTML converter that transforms posts into standalone pages - Blog aggregator that automatically adds new posts to the blog listing - Git automation that deploys posts to production on file save
The workflow integrates with existing infrastructure: a local file watcher detects changes, commits them to GitHub, GitHub Actions triggers, and Namecheap pulls the latest code.
How We Did It
We took a static HTML approach (no server, no database). Each post is:
1. Written as markdown in `/posts/` directory 2. Converted to a standalone HTML page via Python script 3. Automatically added to blog.html's post listing 4. Deployed via existing git→GitHub→Namecheap pipeline
The converter handles: - YAML frontmatter extraction (title, date, tag, excerpt, slug) - Basic markdown-to-HTML translation (headings, bold, italic) - Post card generation with proper tag filtering - Blog listing updates in one command
Why It Matters
This system bridges the gap between AI-generated insights and public showcases. Now we can:
- Document projects in real-time as we build them - Share capabilities without revealing proprietary details - Build credibility by showing actual work and approach - Reduce friction (one `python3 create_post.py file.md` and it's live)
The approach scales from prototype to portfolio while keeping infrastructure minimal and deployment instant.