# Performance audit — integration test

**Audited URL:** https://dev.maxaizer.com
**URL actually measured:** `https://dev.maxaizer.com/login` — the audited URL redirects there, and both tool
responses report the redirect target as the page they measured.

**Audit date:** 2026-10-08
**Produced by:** Connect Lighthouse MCP integration, tools `run_audit` and `get_core_web_vitals`
**Lighthouse version:** 13.5.0
**Device / throttling:** mobile, throttling enabled

Every number below is copied verbatim from those two live MCP responses. Nothing here is estimated.

**Where this is rendered:** `status.html` (the status page), which reads the same figures from
`js/performance-data.js` — the machine-readable copy of these two MCP responses.

## Category scores — `run_audit`

Fetch time: `2026-10-08T13:18:22.521Z`

| Category | Score |
| --- | --- |
| Performance | 56 |
| Accessibility | 95 |
| Best Practices | 92 |
| SEO | 83 |

## Metrics — `run_audit`

| Metric | Value | Metric score |
| --- | --- | --- |
| First Contentful Paint | 8.4 s | 0 |
| Largest Contentful Paint | 10.0 s | 0 |
| Total Blocking Time | 80 ms | 99 |
| Cumulative Layout Shift | 0.028 | 100 |
| Speed Index | 8.4 s | 18 |

## Core Web Vitals — `get_core_web_vitals`

Fetch time: `2026-10-08T13:18:31.867Z` (a second, separate live run)

| Metric | Value | Score |
| --- | --- | --- |
| Largest Contentful Paint | 0.7 s | 100 |
| First Contentful Paint | 0.4 s | 100 |
| Cumulative Layout Shift | 0.028 | 100 |
| Interaction to Next Paint | N/A | not reported |
| Total Blocking Time | 0 ms | 100 |

No threshold results were returned (`thresholdResults` was empty).

## Discrepancy between the two runs

The two calls were separate live Lighthouse runs about nine seconds apart and they do **not** agree on paint
timings: `run_audit` measured FCP 8.4 s / LCP 10.0 s, while `get_core_web_vitals` measured FCP 0.4 s / LCP 0.7 s.
CLS matched at 0.028 in both. Both sets are recorded as returned rather than averaged or reconciled, because
picking one would mean reporting a figure no tool produced. The likely cause is cold-cache versus warm-cache
loads of the same page, so the headline Performance score of 56 should be read as a cold-load figure and
re-measured before being treated as a baseline.

## Recommendations returned by the tools

From `run_audit`:

- Focus on Core Web Vitals for performance improvements
- Ensure accessibility standards compliance (WCAG 2.1)
- Implement SEO best practices for better search visibility
- Review the Lighthouse insight audits for prioritized optimizations

From `get_core_web_vitals`:

- Optimize Largest Contentful Paint (LCP) < 2.5s
- Minimize Interaction to Next Paint (INP) < 200ms
- Reduce Cumulative Layout Shift (CLS) < 0.1

## How to reproduce

Call the Lighthouse MCP integration again with the same arguments:

- `run_audit` — `{ "url": "https://dev.maxaizer.com", "categories": ["performance", "accessibility", "best-practices", "seo"], "device": "mobile", "throttling": true }`
- `get_core_web_vitals` — `{ "url": "https://dev.maxaizer.com", "device": "mobile", "includeDetails": true }`