From Need to Solution: The coffeecup_tech Plugin
🎬 Previously, in Episode 1...
In Part 1, I explained why the usual approaches for managing structured data in Docusaurus eventually reach their limits: duplication, fragile maintenance, and complex multilingual setups.
Even with best practices such as @graph, the core problem remained: structured data lived next to the content instead of being a natural extension of it.
In this Part 2, we move from diagnosis to solution.
I'll show how this reflection led me to design a Docusaurus plugin, built from the ground up to:
- rely on front matter as a single source of truth,
- automatically generate JSON-LD compliant with schema.org,
- handle multilingual content cleanly,
- and remain maintainable over time.