Partial audit: Jev judgments unavailable, so content quality was not assessed; PageSpeed Insights unavailable, so performance used crawl timings only. Scores are an internal rubric for prioritising work, not a prediction of rankings or traffic. Method, formulas and sources are in the appendix.
Executive summary
What matters most on removebg.xyz
removebg.xyz scores 97 out of 100 (grade A) across 7 scored areas. The strongest areas are links and architecture, AI search readiness, performance; the weakest are crawl and indexing, on-page, security and trust. The audit produced 9 actions, 1 of them marked fix first and 2 quick wins.
97/100
overall score, grade A
1
fix-first actions
2
quick wins
0%
Jev answers decisive
JEV-001 · P1rule
Pages excluded from search with noindex
Impact 100 · Hours · 1 page
JEV-002 · P2rule
Duplicate titles across pages
Impact 55 · Hours · 10 pages
JEV-003 · P2rule
Pages with very little main content
Impact 51 · Several days · 2 pages
Executive summary · the plan
What works, what holds it back, what to do
What is working
Links and architecture scores 100.
AI search readiness scores 100.
Performance scores 99.
What is holding the site back
JEV-001: Pages excluded from search with noindex (1 page carry noindex)
Plan
This week
JEV-001 Pages excluded from search with noindex
JEV-002 Duplicate titles across pages
This month
JEV-003 Pages with very little main content
This quarter
JEV-004 No Strict-Transport-Security header
JEV-005 Common security headers missing
JEV-006 Open Graph title or image missing
JEV-007 Pages without a canonical link
JEV-008 Images without width and height
Written by: Automatic summary (no lead-agent narrative was written). A lead agent can replace this with a written narrative by adding narrative.json and re-rendering.
How this audit was made
A source finds, code decides, Jev judges, Claude writes
Each stage does only what it is good at. Anything a count, a status code or a string match can decide is decided by code. Jev, TypeSafe's System One model, answers narrow typed questions about meaning, with probabilities. Nothing is invented to fill a gap: missing data is shown as missing.
1
Crawl
Live fetch from the homepage: robots.txt, sitemaps, internal links, JavaScript rendering when needed.
56URLs · 110 requests
›
2
Rules
52 deterministic checks tied to Google Search Central and web standards.
11rule findings
›
3
Jev judges
Typed Choice, Score and Noul questions, batched per page, with probabilities kept.
0judgments · 0 requests · $0.0000
›
4
Measure
PageSpeed Insights: real-user Core Web Vitals and Lighthouse lab scores.
0PageSpeed runs
›
5
Score and write
Code scores and ranks every action; the narrative is written from the evidence.
9ranked actions
Who decides what
Stage
Owns
Never does
Code
Crawling, status codes, counts, duplicates, schema parsing, formulas, ranking
Guess at meaning
Jev
Page type, search intent, importance, helpfulness, specificity, trust, citability, title and meta fit, competing page pairs
Predict rankings or traffic, invent data, grant permission to act
PageSpeed
Chrome UX Report field data and Lighthouse lab scores
Stand in for every page or device
Claude
The summary and plan, citing action IDs
Add findings that the evidence does not show
What went through each stage
Scorecard
9 areas, one score
Partial audit: Jev judgments unavailable, so content quality was not assessed; PageSpeed Insights unavailable, so performance used crawl timings only. The overall score covers only the areas that were assessed.
Area
Score
Weight
Actions
How it is scored
Crawl and indexing
93
20
2
Rules only
On-page
96
15
1
Rules only
Content quality
n/a
20
1
Not assessed: Jev judgments unavailable
Links and architecture
100
10
0
Rules only
Structured data and sharing
98
8
2
Rules only
AI search readiness
100
12
0
Rules only: Jev judgments unavailable; editorial heuristics, since Google states no special optimization is required for AI features
Impact combines severity, how many pages are affected and how important Jev judged those pages to be. Effort is a planning estimate, not a quote.
Each dot is one action; labelled dots are fix-first or high impact. Colour shows priority.
ID
Action
Priority
Impact
Effort
Pages
By
JEV-001
Pages excluded from search with noindex quick win
P1 Fix first
100
Hours
1
rule
JEV-002
Duplicate titles across pages quick win
P2 Plan next
55
Hours
10
rule
JEV-003
Pages with very little main content
P2 Plan next
51
Several days
2
rule
JEV-004
No Strict-Transport-Security header
P3 When convenient
27
Hours
1
rule
JEV-005
Common security headers missing
P3 When convenient
27
Hours
1
rule
JEV-006
Open Graph title or image missing
P3 When convenient
18
Hours
8
rule
JEV-007
Pages without a canonical link
P3 When convenient
17
Hours
1
rule
JEV-008
Images without width and height
P3 When convenient
17
Hours
2
rule
JEV-009
Structured data missing properties Google requires for rich results
P3 When convenient
17
Hours
2
rule
Site structure
How the site is linked
Pages laid out by clicks from the homepage. Bigger dots are pages Jev judged more important; the outer dashed ring holds pages that no crawled page links to.
Findings · Crawl and indexing
Crawl and indexing 93
HTTP status of crawled URLs
Click depth from the homepage
robots.txt
found · Googlebot allowed · Bingbot allowed
XML sitemap
55 URLs in 1 file(s)
HTTPS
yes · HTTP redirects to HTTPS: yes
Missing page status
404 correct
Rendering
auto; homepage needed JavaScript rendering: no
Crawl limits
60 pages, depth 5
Actions
JEV-001 · Pages excluded from search with noindex
P1highrulequick win
1 affected · 1 page carry noindex
Fix: Confirm each noindex is intended; remove it from pages that should rank. Source
/account
JEV-007 · Pages without a canonical link
P3lowrule
1 affected · 1 of 56 pages
Fix: Add a self-referencing rel=canonical to each indexable page. Source
/account
Findings · On-page
On-page 96
Title length. The 65 mark is a display convention, not a rule.
Main content words per page. The 150 mark is a warning sign, not a ranking factor.
Actions
JEV-002 · Duplicate titles across pages
P2mediumrulequick win
10 affected · 5 titles shared by several pages
Fix: Give each page a title that distinguishes it from the others. Source
Area score starts at 100 and subtracts, per finding, a severity amount (critical 25, high 12, medium 6, low 2, info 0) times (0.5 + 0.5 × share of pages affected). Site-wide findings count as full share. Content quality and AI search readiness blend 70% importance-weighted Jev judgments with 30% rules. Performance blends 50% Lighthouse mobile performance with 50% crawl observations.
Overall score is the weighted mean of scored areas (weights in the scorecard). Unscored areas are left out, never counted as zero. A blocked site is capped at 20; a site without reliable HTTPS at 60.
Impact = severity weight (critical 10, high 6, medium 3, low 1) × (0.6 + 0.4 × reach) × (0.6 + 0.8 × highest Jev importance among affected pages), scaled so the largest is 100. Priority: P1 for critical, or high with impact 40 or more; P2 for impact 30 or more or any high; P3 otherwise. Quick win: impact 35 or more with effort of hours.
Jev
Model requested jev-latest, returned n/a. 0 requests, 0 input tokens, 0 failed, 0 skipped by the budget cap. Cost $0.0000 at $0.042 per million input tokens (output is free). Page text sent to Jev is capped at 6,000 characters. Jev answers are model judgments, not measurements; decisive does not mean correct.
Limits
The crawl covered 56 URLs. Orphan and depth findings only see links on crawled pages.
No Search Console, analytics, backlink or keyword data was used (run with --full for DataForSEO). Nothing here measures rankings, traffic or revenue.
PageSpeed lab scores vary between runs. Field data exists only for sites with enough Chrome traffic.
Rules marked heuristic are editorial conventions, not search engine requirements.