SEO Growth Audit

Built right.
Now give search something to rank.

cadvex.com is one of the cleanest software sites we have crawled: all 72 URLs return 200 with correct canonicals, and AI crawlers can read every page. What is missing is volume and proof. There are 3 articles and 5 short comparison pages, and not one customer is named. AI assistants named CadVex in 2 of 7 buyer questions we asked. Competitors with 151 to 370 guide pages took the other answers.

Prepared byTheStacc
DateNovember 2026
Pages Audited72 crawled · 3 competitors
Overall Score62 / 100

Buyers ask ten questions.
CadVex fully answers three.

We listed the questions a shop owner or estimator asks before buying quoting software. The sources were your own product pages, the comparison tables you publish, competitor pages and live AI answers. Then we matched each question to the best CadVex page. A thin page exists but is too short, too buried or written for a different question. A missing question has no page at all.

CadVex
3
3 blog posts
CloudNC
151
151 blog URLs
Paperless Parts
254
254 blog URLs
QuotedPart
370
370 guide URLs

Editorial pages counted in each site's XML sitemap on 5 November 2026: /blog/ URLs for CadVex, CloudNC and Paperless Parts, /guides/ URLs for QuotedPart. Your 12 user-guide pages are counted separately because they explain the product to people who already use it. Third-party

72
Sitemap URLs (EN + FR)
5
Comparison pages
3
Blog posts
0
Customer proof pages
Buyer questionIntentCadVex pageCoveragePage that answers it today
best CNC machining quoting software for a small job shopcommercial/ (homepage)coveredpaperlessparts.com/cnc-machining/
Paperless Parts alternativecommercial/compare/paperless-partsthinquotedpart.com/paperless-parts-alternative
CloudNC Quote Agent alternativecommercial/compare/cloudnc-quote-agentthinquotedpart.com/cloudnc-alternative
CNC estimating software that runs offline for ITAR / NDA partscommercial/product/desktopcoveredCadVex is the named answer
quote CNC parts from a STEP filecommercial/blog/quoting-from-a-step-file-checklistthincloudnc.com/quote-agent
estimate CNC machining time from a 3D modelinformational/docs/machining-planthinquotedpart.com/guides/calculators/cnc-machining-time
CNC machine hourly rate / quote cost calculatorinformationalNo pagemissingquotedpart.com/guides/calculators/machine-hourly-rate-calculator
DFM analysis software for machine shopscommercial/blog/dfm-checks-that-change-cnc-part-pricethincloudnc.com/quote-agent/dfm
machine shop quoting software with published pricingcommercial/pricingcoveredCadVex is the named answer
quoting software that connects to JobBOSS / ProShop / E2commercialNo pagemissingpaperlessparts.com/blog/not-all-integrations-are-the-same/
3 / 5 / 2
Covered / thin / missing, across 10 buyer questions. The thin pages already exist: 5 comparison pages, the STEP-file checklist and the DFM post. Making those pages longer and linking to them from more of the site is cheaper than writing new pages. The two missing ones are cost calculators and ERP export guides. Competitors use both to win early-stage searches.

AI assistants can read CadVex.
They have little to quote.

Six checks were measured directly on cadvex.com on 5 November 2026. You said ChatGPT sent 5 visitors last week and 2 signed up, so this channel already converts. The score shows which parts are ready and which are holding it back.

VerifiedDirectly observed: page code, robots.txt, sitemap, llms.txt or a live fetch. Artifact on file.
SampledA saved answer from a web-search AI assistant on a stated date. Answers vary between runs, so treat it as directional.
Third-partyCompetitor sitemaps and public search results. Real, but not your own Search Console data.
74out of 100

Ready, but unproven

Crawler access and crawl guidance get full marks, which is rare. You publish llms.txt and llms-full.txt, and all eight AI crawlers we tested got HTTP 200. The score loses most of its points on proof. An assistant has no customer, result or named expert on cadvex.com to quote, so it falls back on sites that have them.

ComponentWeightRaw EvidenceScoreConfidence
AI crawler access20robots.txt allows all public pages; 8 of 8 AI crawler user agents fetched a comparison page with HTTP 20020Verified · high
Entity and business schema20Organization with legalName and 8 sameAs profiles; SoftwareApplication + Offer on both product pages; FAQPage on /faq; comparison pages have BreadcrumbList only17Verified · high
Answer-ready content20/faq 2,334 words and 12 user-guide pages; 3 blog posts; comparison pages 415-478 words13Verified · high
Trust and proof AI can quote150 testimonials, 0 case studies, no named founder or author, /about returns 404; comparison sources dated 28 October 20264Verified · high
Entity consistency15"CadVex" on marketing pages vs "Cadvex" on legal pages; a GetApp listing describes the product as CAD file management10Verified · high
Crawl guidance files10llms.txt and llms-full.txt live; sitemap referenced in robots.txt; no crawl-delay10Verified · high
Composite10020 + 17 + 13 + 4 + 10 + 1074

Who AI names instead

On 5 November 2026 we asked a web-search AI assistant 7 buyer questions that did not mention CadVex. CadVex was named twice: for “best quoting software for a small job shop” and for “estimating software that runs offline for ITAR / NDA parts”. It was absent from the answer to “best alternatives to Paperless Parts”, even though you have a page on exactly that. It was also absent from the answer to “which quoting software publishes its pricing”, though your prices are public. Hosts cited, counted once per answer:

DomainAnswers citing itSource
cloudnc.com6 of 7Sampled
paperlessparts.com5 of 7Sampled
digifabster.com3 of 7Sampled
mtisystems.com3 of 7Sampled
quotedpart.com2 of 7Sampled
cadvex.com2 of 7Sampled

Current SEO health: 62 out of 100

Twelve weighted dimensions, each scored from the crawl of all 72 sitemap URLs, Lighthouse lab runs, competitor sitemaps and the sampled AI answers. Hard caps apply where the evidence requires them. Slow mobile LCP caps Mobile UX at 60, and having no GA4 property caps Analytics at 30.

62out of 100

Strong build, thin library

The technical scores are high, as they should be for a product built by an engineer. The score is held down by everything a new domain has to earn: content volume, customer proof, links into the pages that sell, and mobile speed.

Indexation 90 Structured data 88 Metadata 84 Internal linking 62 Search intent 55 Mobile speed 49 Competitive gap 37 Analytics 30
Indexation and Crawlability90
72/72 sitemap URLs return 200 with self canonicals and index,follow; robots.txt allows public pages and lists the sitemap; the 5 comparison pages have no hreflang pair.
Metadata and SERP Presentation84
Unique titles and descriptions on all 72 URLs, one H1 each, Open Graph complete; /product/desktop title runs 85 chars and two product descriptions exceed 170 chars.
Structured Data and AI Readiness88
JSON-LD types found: Organization, WebSite, ImageObject, VideoObject, SoftwareApplication, Offer, UnitPriceSpecification, QuantitativeValue, FAQPage, Question, Answer, BlogPosting, TechArticle, CreativeWork, BreadcrumbList, ListItem; comparison pages carry BreadcrumbList only.
Performance and Core Web Vitals49
Mobile Lighthouse 51 (lab): LCP 4,658 ms and TBT 1,394 ms are in the poor band; desktop scores 96.
Mobile UX and Accessibility60
Lighthouse accessibility 96 and a clean mobile layout, capped at 60 because mobile LCP is above 4,000 ms.
Content Business Value62
Strong product, pricing, FAQ and docs pages, but 3 blog posts, no customer proof anywhere and comparison pages of 415-478 words.
Search Demand and Intent Coverage55
Of 10 buyer questions, 3 have a strong page, 5 have a thin one and 2 (cost calculators, ERP connections) have none.
Internal Linking and Architecture62
Clear /docs, /blog and /compare hubs; each comparison page and the STEP-file checklist post is linked from only one page.
Conversion Readiness63
"Start free" CTA and published prices on every commercial page; no testimonial or result sits near any CTA; the contact page has no form.
Local Trust and E-E-A-T51
Legal entity, policies, DPA and subprocessor list are thorough; no named founder, author, customer or about page.
Competitive Gap37
3 blog posts against 151 (CloudNC), 254 (Paperless Parts) and 370 guide pages (QuotedPart); 5 comparison pages against QuotedPart's 34 alternative pages; schema depth matches the field.
Analytics and Measurement30
Google Ads tag, PostHog and LinkedIn Insight load only after cookie consent; no GA4 property found, so the score is capped at 30.

The foundation is already done

Most sites we audit need a technical cleanup before content can help. CadVex does not. Keep these as they are and build on them.

Strength

72 of 72 URLs index cleanly, in two languages

Every sitemap URL returns 200 and has a self canonical, index,follow and exactly one H1. No title or description is duplicated. English and French pages are paired with hreflang. The static HTML is prerendered, so Google sees the same title and copy as a browser.

Strength

Built for AI crawlers before most competitors

llms.txt (9,883 bytes) and llms-full.txt are live. All eight AI crawler user agents we tested got HTTP 200. The schema covers 16 distinct JSON-LD types, including SoftwareApplication with Offer pricing on both product pages and FAQPage on a 2,334-word FAQ. CloudNC has no llms.txt at all.

Strength

An honest, sourced comparison format

Each comparison page has a feature table, a “where they are strong” section, dated sources (“as of 28 October 2026”) and a correction email. Assistants prefer to cite pages like this. CadVex is already the named answer for offline / ITAR estimating, the one claim your site states clearly.

Six moves beyond “more pages”

You already have comparison pages and G2 and Capterra profiles, both linked from your llms.txt. These moves start from what exists and fix why it is not working yet.

01
Comparison pages

Make the 5 “vs” pages rankable

Each comparison page is linked from exactly one other page, the /compare hub. They run 415 to 478 words, have no French version and carry BreadcrumbList schema only. Link them from the homepage, both product pages, pricing and every blog post. Expand each to a worked example: the same part quoted in both tools, with screenshots. Add FAQPage and SoftwareApplication schema. Then add one list page, “Paperless Parts alternatives”, which is the exact question where AI left CadVex out. QuotedPart has 34 pages like this.

02
Calculators

Turn the CadVex engine into free calculators

QuotedPart publishes 25 calculators, including a CNC machining-time calculator, a machine hourly rate calculator and a quantity-break calculator. CadVex already computes minutes per feature and the cheapest route by quantity, which is the same maths. One page per calculator targets the shop owner who is not yet looking for software, and each result can end in “quote the whole part in CadVex”.

03
Proof

Name your first customers

No page on cadvex.com names a customer, shows a result or quotes a user, and /about returns 404. Two of the five ChatGPT visitors signed up, so you have users to ask. Three short stories would give assistants something to cite. Each needs the shop, the part, the quote time before and after, and one quote from the estimator. The same users can leave the first reviews on the G2 and Capterra profiles that already exist.

04
Positioning

Own the offline / ITAR niche

“Runs on your computer, nothing uploaded” is the claim AI already credits to CadVex, and no competitor we checked offers it. Build a short cluster around it: quoting ITAR and NDA parts without uploading CAD, Desktop vs cloud data handling, and checklists for customer data agreements. Your /docs/data-and-privacy page is already the technical source for all three.

05
Pricing

Say “published pricing” where buyers look

Asked which quoting software publishes its pricing, the AI assistant named six tools and left out CadVex, though your /pricing page and product schema are public. Write a dated, sourced pricing comparison in the same format as your comparison pages: who publishes prices, who charges per analysis, and who needs a sales call. Put a one-line pricing summary on every comparison page.

06
Listings & entity

Fix how directories describe CadVex

A GetApp snippet describes CadVex as “CAD management software”. GetApp and Capterra share one listing network, so the wrong category can spread to Capterra too. Correct the category to quoting and estimating software, use “CadVex” on the legal pages (they currently say “Cadvex”), and keep the product description identical to llms.txt everywhere.

3 → 60+
Editorial pages, a realistic six-month target. Expanded comparison and alternative pages, calculator guides, the ITAR cluster and customer stories, each mapped to one buyer question. QuotedPart reached 370 guide pages. CadVex does not need that many to be named more often. It needs one strong page per question in the table above.

A short list, because the build is good

These are one-time fixes for whoever ships cadvex.com. None needs a rebuild.

!
Mobile speed is in the poor band — Lighthouse mobile (lab) scores 51: LCP 4,658 ms and total blocking time 1,394 ms. The comparison page scores the same (52; TBT 1,666 ms). Desktop scores 96. The main JavaScript file is 391,303 bytes. Split it per route and load the Paddle, hCaptcha and video scripts only on pages that use them.
!
No GA4 property, and all tracking waits for consent — We found the Google Ads tag (AW-18476902459), PostHog and LinkedIn Insight, all behind a cookie-consent key, and no GA4 ID. The 5 ChatGPT visitors were only those who accepted cookies. Add GA4 with consent mode, verify Search Console, and report chatgpt.com, perplexity.ai and copilot referrals as their own channel.
!
Comparison pages are English-only with breadcrumb schema only — Every other page has an /fr/ pair and hreflang. The /compare pages have neither. Add the French versions with hreflang, plus FAQPage and SoftwareApplication JSON-LD.
!
No author or about page — BlogPosting schema names the company as author, and /about returns 404. Add an about page naming the founder and their machining background. Mark blog posts with a Person author who links to it.
!
Contact page has no form — /contact is 108 words with an email address and no form. A short form with a “send us a part to quote” option gives sales-led buyers a way to start.
!
Three metadata lengths — The /product/desktop title is 85 characters. The /product/cloud and /product/desktop descriptions are 171 and 174 characters. Google truncates all three in results.
✓
Indexation, canonicals, hreflang, robots and sitemap — correct on all 72 sitemap URLs. Keep the static prerender as the site grows.
✓
AI crawler access and llms.txt — all eight AI crawlers we tested got HTTP 200, and llms.txt plus llms-full.txt are live. Add every new page to llms.txt as it ships.

Same buyers, very different libraries

We picked the three domains that came up most often across the 7 AI answers and the sampled web results for the core quoting queries. Every number comes from each site's own sitemap and homepage code, collected on 5 November 2026. Third-party

SignalCadVexPaperless PartsCloudNCQuotedPart
Editorial pages in sitemap3 blog posts254 blog URLs151 blog URLs370 guide URLs
Competitor / alternative pages50 found0 found34
Customer story / case study URLs089210
Calculators00 found0 found25
Homepage JSON-LD types415810
Review / rating schemaNoAggregateRating + ReviewNoNo
llms.txtYes, plus llms-full.txtYes404Redirects (301)
Named in 7 sampled AI answers2562

What's knowable now,
and what isn't

CadVex is a new domain: its sitemap dates begin in September 2026. Some results are certain once the work ships. Others depend on how fast Google trusts a new site. We separate the two.

Links into comparison pages · Deterministic
1 each
5 each
Homepage, both product pages, pricing and related posts. Done the day the template ships.
SEO score · Calculable
62
75
Projection. Assumes GA4, comparison schema, an about page and mobile TBT under 600 ms are shipped. Re-scored with the same 12-dimension method.
AI answers naming CadVex · Directional
2 of 7
Re-test day 60
The same 7 questions, re-asked monthly. Answers vary between runs, so we report the trend, not one result.
Signups from search · Unpredictable
Baseline needed
Tracked from day 1
There is no GA4 or Search Console baseline yet. Set both up first, so signups by channel, AI referrals included, can be measured.
Month 4
Judge it on page-two queries moving up, not page one on day 30. Ranking gains usually take 3-6 months. For a domain this new, the first sign is Search Console queries moving from page two or three to page one, and that comes before signups follow.