Website review, no card needed

wandb.com logo

wandb.com website review

Visit site (opens in a new tab)Reviewed 6 November 202627 issues · 86 checks

This review found 27 issues worth fixing on wandb.com, 8 of them high impact, across 86 checks. The most urgent: render-blocking requests. Your full playbook turns these into prioritised fixes alongside reviews, competitors and channels.

  1. Homepage
  2. Phone render
  3. Links and pages
  4. Google PageSpeed

Google PageSpeed Insights (opens in a new tab), mobile

Real-visitor load time (LCP)75th percentile, phones
9.0 s2.0 sMay 16Oct 312.8 s
Show as table
DateReal-visitor load time (LCP)
May 16, 20268.2 s
May 23, 20268.2 s
May 30, 20268.0 s
Jun 6, 20268.1 s
Jun 13, 20268.1 s
Jun 20, 20267.9 s
Jun 27, 20267.9 s
Jul 4, 20268.0 s
Jul 11, 20268.1 s
Jul 18, 20268.1 s
Jul 25, 20267.9 s
Aug 1, 20267.7 s
Aug 8, 20267.3 s
Aug 15, 20267.0 s
Aug 22, 20266.7 s
Aug 29, 20266.5 s
Sep 5, 20266.6 s
Sep 12, 20266.9 s
Sep 19, 20267.1 s
Sep 26, 20267.2 s
Oct 3, 20267.2 s
Oct 10, 20267.0 s
Oct 17, 20266.6 s
Oct 24, 20265.1 s
Oct 31, 20262.8 s
Real-visitor tap response (INP)75th percentile, phones
574 ms108 msMay 16Oct 31162 ms
Show as table
DateReal-visitor tap response (INP)
May 16, 2026497 ms
May 23, 2026504 ms
May 30, 2026510 ms
Jun 6, 2026520 ms
Jun 13, 2026518 ms
Jun 20, 2026505 ms
Jun 27, 2026503 ms
Jul 4, 2026490 ms
Jul 11, 2026489 ms
Jul 18, 2026485 ms
Jul 25, 2026467 ms
Aug 1, 2026454 ms
Aug 8, 2026440 ms
Aug 15, 2026428 ms
Aug 22, 2026423 ms
Aug 29, 2026422 ms
Sep 5, 2026421 ms
Sep 12, 2026437 ms
Sep 19, 2026444 ms
Sep 26, 2026446 ms
Oct 3, 2026446 ms
Oct 10, 2026429 ms
Oct 17, 2026410 ms
Oct 24, 2026317 ms
Oct 31, 2026162 ms

Chrome UX Report (opens in a new tab): real Chrome visitors over rolling 28-day periods.

One score from every part of this review. How it works

Full-page screenshot of wandb.com (phone) rendered by Google PageSpeed Insights

How it loaded on a phone

  1. wandb.com at 0.4 s0.4 s
  2. wandb.com at 0.8 s0.8 s
  3. wandb.com at 1.1 s1.1 s
  4. wandb.com at 1.5 s1.5 s
  5. wandb.com at 1.9 s1.9 s
  6. wandb.com at 2.3 s2.3 s
  7. wandb.com at 2.6 s2.6 s
  8. wandb.com at 3.0 s3.0 s

Speed

4 to fix80
  • Render-blocking requests

    2.6 s

    30 files block the first paint until they finish downloading. Deferring or inlining them could save about 2.6 s.

  • Mobile performance score: 80

    80/100

    Your mobile performance score is 80. Pages that score under 90 lose visitors before the first paint.

  • Improve image delivery

    259 KiB

    5 images could be 259 KiB smaller with modern formats (WebP or AVIF) and sizes that fit a phone screen.

  • 2,148 HTML elements

    The page has 2,148 HTML elements. Above 1,500, rendering and every interaction get slower on phones.

  • 8 third-party services on the page

    8 third-party services load with your homepage (Google Tag Manager, Cloudflare CDN and coreweave.com and others), adding 459 KiB and 318 ms of main-thread time.

  • Reduce unused CSS

    150 ms · 61 KiB

    About 61 KiB of CSS is downloaded but never applied on this page. Google estimates it would make the page about 150 ms faster on a phone.

  • 4 long tasks on the main thread

    4 long tasks block the main thread while loading, the longest for 135 ms. During a long task the page ignores taps.

  • Reduce unused JavaScript

    110 KiB

    About 110 KiB of JavaScript is downloaded but never used on this page.

  • Use efficient cache lifetimes

    102 KiB

    5 files are served with short cache lifetimes, so returning visitors download them again (102 KiB).

  • Font display

    50 ms

    Web fonts hide text while they load. Setting font-display: swap shows text immediately (about 50 ms).

  • Minify CSS

    7 KiB

    7 KiB could be saved by minifying CSS.

  • Chained critical requests

    Critical files load one after another instead of in parallel (3 chains). Preconnecting to key origins and preloading the hero image shortens the chain.

  • Forced reflow

    Scripts make the browser recalculate the layout repeatedly while the page loads, which wastes processor time on phones.

  • Minify JavaScript

    2 KiB

    2 KiB could be saved by minifying JavaScript.

2 checks passed
  • Main-thread work: 1.9 s

    A phone's processor spends 1.9 s on main-thread work while loading, 781 ms of it running JavaScript. That is comfortably low.

  • Server response: 17 ms

    17 ms

    Your server sent the HTML in 17 ms, well inside Google's 600 ms guideline.

Core Web Vitals

5 to fix100
  • Speed Index: 6.5 s

    6.5 s

    Visually, the page fills in at 6.5 s (Speed Index). Good is under 3.4 s.

  • Time to Interactive: 8.2 s

    8.2 s

    The page becomes fully interactive after 8.2 s. Good is under 3.8 s.

  • First Contentful Paint: 3.1 s

    3.1 s

    The first text or image appears after 3.1 s on a phone. Good is under 1.8 s.

  • Largest Contentful Paint: 3.1 s

    3.1 s

    Your largest element takes 3.1 s to render on a phone. Google's threshold for “good” is 2.5 s. The largest element is the text “import weave weave.init("quickstart") @weave.op() def llm_a…”. Most of its time is element render delay (2.5 s).

  • 22 images without width and height

    22 images have no width and height attributes, so the layout jumps as each one loads.

10 checks passed
  • Passes Core Web Vitals for real visitors

    Real Chrome users on phones pass Google's Core Web Vitals on this page (28-day field data from the Chrome UX Report (opens in a new tab)).

  • Real-user LCP: 2.0 s

    2.0 s

    Real visitors wait 2.0 s for the largest content to appear (75th percentile). Good is under 2.5 s.

  • Real-user INP: 129 ms

    129 ms

    Real visitors wait 129 ms for the page to respond to a tap (75th percentile). Good is under 200 ms.

  • Real-user CLS: 0.01

    0.01

    For real visitors the layout shifts by 0.01 (75th percentile). Good is under 0.1.

  • Real-user FCP: 1.6 s

    1.6 s

    Real visitors see the first content after 1.6 s (75th percentile). Good is under 1.8 s.

  • Real-user TTFB: 706 ms

    706 ms

    Real visitors wait 706 ms for your server to start answering (75th percentile). Good is under 800 ms.

  • Tap response (INP) got 52% faster for real visitors

    -52%

    Over the last 25 four-week periods to 2026-10-31, real Chrome visitors' tap response (inp) went from 497 ms to 162 ms at the 75th percentile (Chrome UX Report (opens in a new tab) history).

  • Largest content (LCP) got 52% faster for real visitors

    -52%

    Over the last 25 four-week periods to 2026-10-31, real Chrome visitors' largest content (lcp) went from 8.2 s to 2.8 s at the 75th percentile (Chrome UX Report (opens in a new tab) history).

  • Total Blocking Time: 164 ms

    164 ms

    Scripts block the page for only 164 ms while loading. Good is under 200 ms.

  • Cumulative Layout Shift: 0

    0

    The layout barely moves while loading (CLS 0, good is under 0.1).

Images and page weight

1 to fix90
  • 2 images far larger than shown

    119 KiB

    2 images are downloaded at several times the size a phone draws them, for example app-advert-2.png at 472 px wide, shown at 109 px. Resized copies would save about 119 KiB.

  • About 0.4 g of CO₂ per visit

    0.39 g

    By the Sustainable Web Design model, a visit to your homepage emits about 0.39 g of CO₂e: 46.3 kg a year at 10,000 visits a month. It scales with page weight, so every byte cut lowers it.

3 checks passed
  • Page weight: 1.4 MB in 122 requests

    1.4 MB

    Your homepage downloads 1.4 MB in 122 requests on a phone: scripts 463 KiB (58), images 416 KiB (24) and fonts 269 KiB (9). That is a lean page.

  • Images use modern formats

    Every sizeable image is WebP, AVIF or SVG.

  • Images load when needed

    Images below the first screen wait until the visitor scrolls to them.

Accessibility

6 to fixGoogle 8992
  • 2 text elements with low contrast

    2 text elements don't have enough contrast against the background to read easily, especially outdoors on a phone.

  • Scrollable region must have keyboard access (1 element)

    Scrollable region must have keyboard access: 1 element on the page fails this check, rated serious by axe-core.

  • 1 tap target too small

    1 tap target is too small or too close together to hit reliably with a finger.

  • Landmarks should have a unique role or role/label/title (i.e. accessible name) combination (1 element)

    Landmarks should have a unique role or role/label/title (i.e. accessible name) combination: 1 element on the page fails this check, rated moderate by axe-core.

  • All page content should be contained by landmarks (1 element)

    All page content should be contained by landmarks: 1 element on the page fails this check, rated moderate by axe-core.

  • 2 links without text

    2 visible links have no text or label, so screen readers cannot say where they go.

  • Headings skip levels

    Headings skip levels (h2 → h5), which makes the page harder to scan with a screen reader.

2 checks passed
  • Every image has alt text

    All 23 images in the HTML carry an alt attribute.

  • 47 automated accessibility rules passed

    axe-core 4.13.0, the open-source engine behind most accessibility testing tools, ran 54 WCAG 2.2 A/AA and best-practice rules on the phone render: 47 passed and 7 failed. 4 mores need a person to judge. Automated rules catch about a third of accessibility problems; the rest need manual testing.

SEO

3 to fixGoogle 9289
  • 2 H1 headlines

    There are 2 <h1> headlines (“AI is easy to productionize” and “Now available:”). One clear main headline is easier for people and search engines.

  • Structured data has 1 problem

    Structured data found (Corporation, WebPage and WebSite) but with 1 problem, for example “WebPage: keywords is an empty list”.

  • 1 link with vague text

    1 link use generic text like “click here” or “learn more”, which tells Google nothing about the page behind it.

  • 267 internal links, 20 external

    The homepage links to 267 pages on wandb.com and 20 external pages.

6 checks passed
  • Meta description: 141 characters

    Your meta description is 141 characters: “Learn why thousands of companies rely on W&B as their system of record for training AI models and developing AI applica…”.

  • Title: 43 characters

    Your title is a healthy 43 characters: “Weights & Biases: The AI developer platform”.

  • Canonical URL set

    The canonical URL is https://wandb.ai/site/.

  • Open Graph tags complete

    Open Graph tags are complete, so shared links show a title, description and image.

  • robots.txt in place

    robots.txt is in place and points crawlers to your sitemap.

  • XML sitemap found

    Your XML sitemap is reachable at https://wandb.ai/sitemap.xml.

AI search readiness

100
  • No llms.txt

    There is no https://wandb.ai/llms.txt, a short Markdown guide to your key pages for AI tools (llmstxt.org). It is optional and few AI tools read it yet, so treat it as a cheap extra.

3 checks passed
  • AI answer engines may read the site

    Your robots.txt lets the crawlers behind ChatGPT search, Claude and Perplexity read the site and cite it in answers.

  • Organization data: 5 of 6 facts

    Your Organization JSON-LD gives name, url, logo, sameAs (social profiles) and description; contact details would complete it.

  • Text is in the HTML

    Your server sends the page's text in the HTML (1,387 words), so crawlers that don't run JavaScript still read it.

Other pages

3 to fix83
  • 2 linked pages hidden from search

    https://wandb.ai/site/solutions/evaluations and https://wandb.ai/site/solutions/quant-trading are marked noindex but linked from your homepage. If they are meant to be found, remove the noindex.

  • 3 pages without a meta description

    https://wandb.ai/site/solutions/evaluations, https://wandb.ai/site/solutions/rag and https://wandb.ai/site/solutions/quant-trading have no meta description, so Google writes the search snippet from whatever text it finds.

  • 2 links to non-canonical URLs

    Your homepage links to 2 pages by a URL that isn't their canonical one, for example https://wandb.ai/site/solutions/ai-governance, whose canonical is https://wandb.ai/site/solutions/rag. Linking straight to the canonical URL saves crawlers a detour and keeps ranking signals on one address.

  • 15 internal pages checked

    We fetched 15 pages linked from your homepage (of 72 found) and read each the way a search engine does.

4 checks passed
  • Linked pages load

    All 15 pages we checked loaded without an error.

  • Titles are unique

    Every page we checked has its own title.

  • Every page has an H1

    Each page we checked has a main headline.

  • Every page has a title

    All 15 pages that loaded have a <title>.

Best practices

2 to fixGoogle 9687
  • 2 small tap targets

    2 of 7 tappable elements near the top of the page are smaller than 24 px, hard to hit with a thumb.

  • 3 JavaScript errors while loading

    3 errors were logged to the browser console while the page loaded, for example “Failed to load resource: the server responded with a status of 404 ()”.

1 checks passed
  • Mobile viewport set

    The page declares a mobile viewport.

Content and conversion

3 to fix71
  • No social proof

    No testimonials, customer logos, ratings or customer counts on the homepage: nothing tells a new visitor that others already trust you.

  • Local business without LocalBusiness data

    Your homepage shows a street address but no LocalBusiness JSON-LD, so Google and AI assistants have to guess your address, hours and phone from the page text.

  • No phone number on the homepage

    A local customer looking to call finds no phone number on the homepage. Add one, ideally as a tap-to-call link.

  • No opening hours

    The homepage shows no opening hours. If customers visit or call, show when.

  • No form on the homepage

    There is no form on the homepage, so every lead has to click through to another page first.

  • No map or directions link

    There is no link to Google or Apple Maps. A directions link is one tap for a phone visitor.

  • No risk reducer

    No guarantee or risk reducer (money-back, free trial, no credit card) is mentioned, which is often what tips a hesitant visitor.

  • About 409 words of copy

    The homepage carries about 409 words of visible text.

4 checks passed
  • Clear next step on the first screen

    On a phone, the first screen shows a clear next step: “GET STARTED WITH WEAVE” and “REQUEST DEMO”.

  • 11 calls to action

    The homepage has 11 calls to action, for example “Contact centers”, “Pricing” and “Contact”.

  • Analytics in place

    Analytics detected: Google Tag Manager.

  • Direct contact links

    Visitors can reach you directly: 0 phone links and 1 email link.

Privacy and tracking

  • 1 tracker load before any consent

    1

    Before a visitor clicks anything, the page loads Google Tag Manager (measured from a US location before any click). A consent tool (OneTrust) is installed; check it blocks these until visitors agree where the law requires it.

  • 2 cookies set on arrival

    2

    Opening the homepage sets 2 cookies (measured from a US location before any click).

  • 13 other companies contacted

    Loading the homepage contacts 13 hosts run by other companies (fonts, scripts, analytics, ads), each of which sees the visitor's IP address.

1 checks passed
  • Consent tool: OneTrust

    A consent management tool (OneTrust) loads with the page.

Security, domain, email and software versions

The security, domain and email checks, software and plugin versions, and any app with a known security problem are visible only to people at wandb.com, because details like these could help someone attack a site. Every review keeps them private, whatever they found.

Upkeep, stack and company

Stack

CMS
WordPress
Framework
jQuery
Analytics
Google Tag Manager | Segment
Hosting
Squarespace

Company

Headquarters
San Francisco

Upkeep, stack, hiring and company data from programmable.info

Go beyond the website

What your $25 Growth Playbook adds

This review found 27 issues on wandb.com, including 9 speed issues, 3 SEO issues and 3 conversion gaps. That is the website alone. The full Growth Playbook pairs these findings with what your customers say in public reviews and communities and what your closest competitors do better, then turns everything into at least 20 prioritised ideas, each with the evidence behind it, a step-by-step playbook, a metric and an owner.

  • 28 research agents read your site, reviews, competitors, communities and public filings
  • Customer review mining (G2, Capterra, Trustpilot and more) and voice-of-pain analysis in your buyers’ own words
  • Competitor intelligence with battlecards for your closest rivals
  • Ideal customer profile, buyer segments and buying-stage funnel
  • Ranked channels, outreach sequences and content angles
  • Live Reddit threads where buyers ask for what you sell, with draft replies
  • Earned media and PR openings
  • At least 20 ideas with evidence, a step-by-step playbook, metric, owner, effort and impact, grouped into this week, this quarter, big bets and wildcards, plus a 30/60/90-day roadmap
  • PDF, Markdown, CSV, Notion and Linear exports, and a shareable page your team can track progress on
  • Refresh the playbook later to see what changed

By continuing you agree to the Terms and acknowledge the Privacy Policy.

Free requests are written in the order they come in, so it can take a while. Pay $25 and yours is ready within a few hours. Either way you get a shareable link and a PDF.

  • No subscription
  • Free, or $25 to skip the line
  • Money-back guarantee

14-day money-back guarantee: not happy for any reason, email us within 14 days and we refund in full.

Get the full playbook · $25