Performance: pre-load post meta in a single query on the pages list endpoint, eliminating N+1 database queries.
1.0.4
Performance: decrypted API key is now cached in memory for the duration of the request, avoiding redundant database reads and OpenSSL decryption on repeated calls.
Added cache invalidation on key update and deletion for consistency.
1.0.2
Performance optimization: prompt templates are now cached in-memory per request during generation.
Reduced unnecessary database writes by avoiding model option updates during each generation call.
Minor internal cleanup of OpenAI response handling to avoid redundant body parsing.
1.0.1
Fixed API responses to always return valid JSON
Improved error handling in frontend API helper
Prevented HTML leaks during content generation
Added content length validation before AI processing
Enhanced robustness against shortcode interference
Better user feedback for generation failures
1.0.0
Initial public release.
AI generation (OpenAI) & injection of FAQ / HowTo / Event / Product JSON-LD.
