This case documents a Nexora Engine validation run on a WordPress site built with Elementor Pro, hosted on a single VPS (2 vCPU, 4GB RAM, Apache + PHP-FPM). The goal was to establish whether Engine’s static mirror approach could reduce TTFB and improve LCP on an Elementor site without disabling Elementor’s built-in asset loading mechanisms.
Environment
- WordPress 6.5, Elementor Pro 3.21
- Apache 2.4 on Ubuntu 22.04, PHP 8.2-FPM
- WooCommerce active (shop pages excluded from mirroring)
- Existing cache plugin: LiteSpeed Cache (disabled for Engine comparison)
- No CDN in baseline configuration
Baseline measurements
Measurements were taken using WebPageTest (Dulles, Virginia, Chrome, cable throttling) with a five-run average. Baseline LiteSpeed Cache configuration was in place for baseline runs.
- TTFB: 680ms (median)
- LCP: 2.8s (hero image, Elementor widget)
- CLS: 0.04
- Total page payload: 2.1MB
Engine configuration
Engine was configured with a 6-hour mirror freshness window for static pages (home, about, services, blog) and excluded WooCommerce product, cart, and checkout URLs. The Apache rewrite rules were deployed to the virtual host configuration rather than .htaccess to avoid per-request overhead.
Elementor’s CSS print method was left on “Internal CSS” (the default) — Engine captures the rendered HTML including inline styles, so the static mirror includes Elementor’s generated CSS without requiring any plugin configuration changes.
Post-Engine measurements
- TTFB: 48ms (median) — 93% reduction
- LCP: 1.6s — 43% improvement (remaining LCP time is image fetch, not server)
- CLS: 0.04 — unchanged
- Total page payload: 2.1MB — unchanged (Engine does not modify assets, only delivery)
Observations
The TTFB reduction was the most significant result. Serving pre-rendered HTML from disk eliminated PHP, MySQL, and Elementor’s template rendering from the critical path for cached pages. The remaining LCP time is dominated by the hero image download, which is outside Engine’s scope (addressed separately by Nexora Media).
CLS was unaffected because Engine captures the fully rendered DOM including Elementor’s layout calculations. There were no layout shift regressions from the static mirror.
WooCommerce exclusions worked as expected — dynamic pages fell through to PHP normally without any cache poisoning issues.
This validation scenario is available as a demo request for Elementor sites on Apache or Nginx. Include your builder and server configuration in the request form.