Skip to content
PERFORMANCE LAB / FREE AUDIT

Run a free image audit on any public URL. Get a 12-point performance report in 20 minutes.

Pixlog's Performance Lab crawls your site from 47 edge PoPs, scores every product and hero image against Lighthouse, Web Vitals and Core Image metrics, and ships a written diagnostic to your inbox. No login. No credit card. No sales call.

NO CREDIT CARD/AUDIT IN 20 MIN/SOC 2 TYPE II

9.2Bimage transformations / day
2,800+brands audited
47edge PoPs
pixlog@lab:~/audit
$ pixlog audit --url https://yourstore.com
→ crawling from 47 PoPs
→ indexing 1,284 image assets
→ scoring: LCP · CLS · format coverage · payload
[✓] report ready · 12 checkpoints · 4-7 min
BEFORE
total image payload14.8 MB
LCP image4.2 s
lighthouse perf52
AFTER (PROJECTED)
total image payload2.1 MB
LCP image1.4 s
lighthouse perf98
// AUDIT OUTPUT · SAMPLE FROM Q2 2024
12-CHECKPOINT DIAGNOSTIC

What the audit measures — the 12 image checkpoints your team actually argues about.

Each run crawls every public image on the submitted URL and returns a scored report across four diagnostic categories. No estimates, no modeled data — real numbers from real edge fetches.

01

Page weight & payload

Total image bytes shipped per page, broken down by hero, product grid, thumbnail and below-fold assets. The single number that moves Lighthouse more than any other.

  • total_payload_kb/per_image_mean_kb
  • per_format_breakdown/above_fold_ratio
02

Format & codec coverage

Detection of AVIF, WebP, JPEG XL and HEIC delivery per visitor device — plus how many images are still served as legacy JPEG or PNG where a modern codec would shrink payload 40-65%.

  • avif_coverage_pct/webp_coverage_pct
  • legacy_jpeg_count/oversized_png_count
03

Core Web Vitals & LCP

Largest Contentful Paint timing per page, with the offending hero or product image isolated. Includes CLS from images without explicit dimensions and the cumulative layout shift contribution.

  • lcp_ms/cls_score
  • image_lcp_ratio/missing_dim_count
04

Delivery & caching

Time-to-first-byte from each PoP, cache-hit ratio on returning assets, and whether the origin is shipping with stale or missing Cache-Control headers that force revalidation.

  • median_ttfb_ms/p95_ttfb_ms
  • cache_hit_ratio/missing_max_age_count
BEFORE / AFTER · ANONYMIZED

Three real sites. Hard numbers. Zero modeling.

Each pair below is a production site audited and deployed in the last 90 days. Names redacted under customer NDA. Payload, LCP and Lighthouse deltas are post-deploy, measured by the customer's own RUM vendor.

CASE 01 · DTC APPAREL fashion-brand.eu
Before optimization: heavy product imagery
BEFORE
PAYLOAD11.4 MB
LCP3.8 s
LIGHTHOUSE48
After optimization: crisp optimized product imagery
AFTER · PIXLOG
PAYLOAD1.8 MB
LCP1.3 s
LIGHTHOUSE97
−84% PAYLOAD −66% LCP +49 LIGHTHOUSE
CASE 02 · HOME FURNISHINGS homegoods-marketplace.com
Before optimization: heavy lifestyle imagery
BEFORE
PAYLOAD18.2 MB
LCP4.6 s
LIGHTHOUSE41
After optimization: crisp optimized lifestyle imagery
AFTER · PIXLOG
PAYLOAD3.1 MB
LCP1.7 s
LIGHTHOUSE94
−83% PAYLOAD −63% LCP +53 LIGHTHOUSE
CASE 03 · BEAUTY BRAND skincare-direct.co
Before optimization: heavy hero product imagery
BEFORE
PAYLOAD7.9 MB
LCP3.1 s
LIGHTHOUSE61
After optimization: crisp optimized hero imagery
AFTER · PIXLOG
PAYLOAD1.4 MB
LCP1.1 s
LIGHTHOUSE99
−82% PAYLOAD −64% LCP +38 LIGHTHOUSE
INDEPENDENTLY VERIFIED

The benchmarks performance teams cite internally.

FORRESTER WAVE · 2024
4.7/5

Named a Leader

Image & Media Optimization Wave, scoring 4.7/5 on time-to-value across the eight vendors evaluated.

G2 GRID · SPRING 2024
#1RANKED

Top of Image Optimization

412 verified reviews averaging 4.8/5 — the highest satisfaction score in the category.

CATCHPOINT · Q2 2024
71%MEDIAN CUT

Independent payload benchmark

Median image payload reduction within 10 minutes of install — verified across 312 production sites.

WEBBY AWARDS · 2024
WINNERTECH ACHIEVEMENT

Best Technical Achievement

Recognized in the E-commerce category for the per-device format selection pipeline.

METHODOLOGY

How the audit actually runs.

The technical questions your team will ask before submitting a URL — answered up front, no sales call required.

Q → Where does the crawl run from?

From all 47 Pixlog edge PoPs in parallel — Frankfurt, São Paulo, Singapore, Sydney, Virginia, Mumbai and 41 others. The report shows the median and p95 numbers across PoPs, not a single-region snapshot.

Q → How long does it take?

4 to 7 minutes for most sites. The crawler fetches every public image once per PoP, runs format detection, and assembles the 12-checkpoint score. You'll get an email the moment it's ready.

Q → Do I need to log in or share credentials?

No login, no OAuth, no install. Submit a public URL and the crawler only fetches what an unauthenticated visitor would see. Authenticated product imagery is not included in the audit by default.

Q → What happens to the data?

Audit results are stored against your email for 30 days, then purged. We do not sell, share, or enrich the data with third-party intent providers. SOC 2 Type II audited since November 2023.

FREE · NO LOGIN · 20 MIN

Run the audit. See your numbers.

Paste a public URL. The crawler does the rest — payload, LCP, Lighthouse, format coverage, all 12 checkpoints, in your inbox before your next standup.

https://
NO CREDIT CARD RESULTS IN 20 MIN SOC 2 TYPE II
PIXLOG PERFORMANCE LAB · LISBON / SAN FRANCISCO / SINGAPORE