Accessibility
· 13 checks — Landmarks, headings, alt text, forms, and link quality rolled into one auditable list.DLink & Button QualityAction6 issue(s) across 141 links and 16 buttonsFIX
| Element | Text | Issue | Suggested Fix |
|---|---|---|---|
| https://policies.google.com/technologies… | Learn more about how Google us… | new tab | Add '(opens in new tab)' to text |
| https://colab.research.google.com/github… | Run quickstart | new tab | Add '(opens in new tab)' to text |
| https://blog.tensorflow.org/2025/08/what… | (empty) | empty | Add link text or aria-label |
| https://blog.tensorflow.org/2025/03/what… | (empty) | empty | Add link text or aria-label |
| https://blog.tensorflow.org/2024/12/intr… | (empty) | empty | Add link text or aria-label |
| https://blog.tensorflow.org/2024/11/mlsy… | (empty) | empty | Add link text or aria-label |
| https://blog.tensorflow.org/2024/10/what… | (empty) | empty | Add link text or aria-label |
| https://blog.tensorflow.org/2024/07/what… | (empty) | empty | Add link text or aria-label |
| /resources/models-datasets | ML models & datasets … | img no alt | Add alt attribute to the image |
| /resources/libraries-extensions | Libraries & extensions … | img no alt | Add alt attribute to the image |
| /resources/tools | Developer tools … | img no alt | Add alt attribute to the image |
Links without text are announced as raw URLs by screen readers.
https://blog.tensorflow.org/2025/08/whats-new-in-tensorflow-2-20.html; https://blog.tensorflow.org/2025/03/whats-new-in-tensorflow-2-19.html; https://blog.tensorflow.org/2024/12/introducing-wake-vision-new-dataset-for-p…; https://blog.tensorflow.org/2024/11/mlsysbookai-principles-and-practices-of-m…; https://blog.tensorflow.org/2024/10/whats-new-in-tensorflow-218.html; https://blog.tensorflow.org/2024/07/whats-new-in-tensorflow-217.html
Links with no accessible text (empty <a></a>, image-only no alt, icon-only no aria-label) are unidentifiable to screen readers.
Source: WCAG 2.1 SC 2.4.4
Image-only links need alt text on the image (or aria-label on the link) so screen readers can announce them.
a[href="/resources/models-datasets"]; a[href="/resources/libraries-extensions"]; a[href="/resources/tools"]
Image-only links with no alt are unidentifiable to screen-reader users — link's destination is invisible.
Source: WCAG 2.1 SC 2.4.4
Add '(opens in new tab)' to link text or aria-label.
https://policies.google.com/technologies/cookies?hl=en; https://colab.research.google.com/github/tensorflow/docs/blob/master/site/en/…
Links with target="_blank" without rel="noopener" leak the originating page's window context — security and UX issue.
Learn more ▾ ▴
Without rel="noopener", the new tab can navigate the original tab via window.opener (tab-nabbing attack). Modern browsers default to noopener for target=_blank but only since recent versions. Always set rel="noopener noreferrer" explicitly.
Source: MDN target / OWASP
DPrint StylesheetActionNo print stylesFIX
BHeading Hierarchy25 headingsREVIEW
- H2 Get started with TensorFlow
- H2 Solve real-world problems with ML
- H2 What's new in TensorFlow
- H3 What's new in TensorFlow 2.20
- H3 What's new in TensorFlow 2.19
- H3 Introducing Wake Vision: A High-Quality, Large-Scale Dataset for TinyML Computer Vision Applications
- H3 MLSysBook.AI: Principles and Practices of Machine Learning Systems Engineering
- H3 What's new in TensorFlow 2.18
- H3 What's new in TensorFlow 2.17
- H2 Explore the ecosystem
- H3 TensorFlow.js
- H3 LiteRT
- H3 tf.data
- H3 TFX
- H3 tf.keras
- H3 Kaggle Models
- H3 TensorFlow Datasets
- H3 TensorBoard
- H3 ML models & datasets
- H3 Libraries & extensions
- H3 Developer tools
- H2 Stay connected
- H2 Start building with TensorFlow
- H3 Stay connected
- H3 Support
Every page should have one H1 that describes the page content.
No H1 means screen-reader users can't identify the page's primary topic, and Google's content-extraction degrades.
Learn more ▾ ▴
The H1 is the document title for assistive tech and a strong signal to search engines about page topic. Pages without one force screen readers to fall back to the <title> attribute or page chrome. Add a single H1 that names the page's primary subject.
Source: WCAG 2.4.6 / Google Search Central
BAlt Text Quality3 of 26 images have issuesREVIEW
| Issue | Count |
|---|---|
| missing | 3 image(s) |
Images without alt text are invisible to screen readers.
Each image without alt text is a WCAG 1.1.1 failure — invisible to screen-reader users, lost from Google Image Search.
Learn more ▾ ▴
WCAG 2.1 Level A requires text alternatives for non-decorative images. Empty alt='' is fine for decorative; meaningful images need descriptive text. Common fixes: CMS audit + bulk add, build-time linter (alt-text-required ESLint rule), CI gate on Lighthouse a11y score.
Source: WCAG 2.1 SC 1.1.1 / WebAIM Million Report
B404 Error PageHTTP 404, custom pageREVIEW
BFavicon & Branding12 icon(s) detectedREVIEW
BDark Mode SupportDark mode detectedREVIEW
Detection limited to meta tags and inline styles.
BLighthouse Accessibility AuditsScore 84/100 — 4 failing, 26 passedREVIEW
Accessibility
These checks highlight opportunities to improve the accessibility of your web app. Automatic detection can only detect a subset of issues and does not guarantee the accessibility of your web app, so manual testing is also encouraged.
Contrast
Low-contrast text is difficult or impossible for many users to read. Learn how to provide sufficient color contrast.
Performance issues directly impact user engagement and conversion rates.
| Failing Elements |
|---|
Updated August 19, 2025 div.devsite-card-content-wrapper > div.devsite-card-content > div.devsite-card-section > p.devsite-card-date |
Updated March 13, 2025 div.devsite-card-content-wrapper > div.devsite-card-content > div.devsite-card-section > p.devsite-card-date |
LIBRARY li > a > div.tensorsite-blog-carousel__wrapper > p.tensorsite-blog-carousel__eyebrow |
LIBRARY li > a > div.tensorsite-blog-carousel__wrapper > p.tensorsite-blog-carousel__eyebrow |
These are opportunities to improve the legibility of your content.
Names and labels
Informative elements should aim for short, descriptive alternate text. Decorative elements can be ignored with an empty alt attribute. Learn more about the `alt` attribute.
Performance issues directly impact user engagement and conversion rates.
| Failing Elements |
|---|
div.tensorsite-container > div.tensorsite-adjacent-partials-wrapper > div.tensorsite-adjacent-partials__main-section > img div.tensorsite-container > div.tensorsite-adjacent-partials-wrapper > div.tensorsite-adjacent-partials__main-section > img |
div.tensorsite-container > div.tensorsite-adjacent-partials-wrapper > div.tensorsite-adjacent-partials__secondary-section > img div.tensorsite-container > div.tensorsite-adjacent-partials-wrapper > div.tensorsite-adjacent-partials__secondary-section > img |
Screen reader users rely on frame titles to describe the contents of frames. Learn more about frame titles.
Performance issues directly impact user engagement and conversion rates.
| Failing Elements |
|---|
div.tensorsite-home-hero__main-section > div.tensorsite-home-hero__header > devsite-iframe > iframe div.tensorsite-home-hero__main-section > div.tensorsite-home-hero__header > devsite-iframe > iframe |
div.tensorsite-stay-connected__container > div.tensorsite-stay-connected__steps-container > devsite-iframe > iframe div.tensorsite-stay-connected__container > div.tensorsite-stay-connected__steps-container > devsite-iframe > iframe |
These are opportunities to improve the semantics of the controls in your application. This may enhance the experience for users of assistive technology, like a screen reader.
Best practices
Disabling zooming is problematic for users with low vision who rely on screen magnification to properly see the contents of a web page. Learn more about the viewport meta tag.
Informational: a Permissions-Policy directive showing feature -> allowed origins.
Source: MDN Permissions-Policy
| Failing Elements |
|---|
head > meta head > meta |
These items highlight common accessibility best practices.
ALandmark Structure10 landmarksPASS
Multiple navigations need aria-label to distinguish them for screen readers.
Some <nav> elements lack aria-label — screen-reader users hear 'navigation' multiple times with no way to distinguish them.
Learn more ▾ ▴
When a page has multiple <nav> regions (primary, footer, breadcrumb), each needs aria-label or aria-labelledby. AT users navigate by landmark; identical 'navigation' announcements force them to enter each one to discover purpose.
Source: WAI-ARIA Authoring Practices
A+Form AccessibilityAll 2 controls labeledPASS
| Control | Type | Label | Method |
|---|---|---|---|
| q | text | Search | aria-label |
| input | text | Type to filter | aria-label |
A+Web ManifestPWA-readyPASS
A+Color Contrast (Screenshot)20 text elements analyzed, 0 fail WCAG AAPASS
Analyzes text contrast against the actual rendered page, including background images, gradients, and overlays that CSS-based tools cannot detect.
Show all checked elements (20)
| Element | Ratio | Required | FG | BG | Result |
|---|---|---|---|---|---|
| h2 Get started with Ten… | 18.33:1 | 3.0:1 | #000000 | #E8F0FE | Pass |
| h2 Solve real-world pro… | 18.33:1 | 3.0:1 | #000000 | #E8F0FE | Pass |
| h2 What's new in Tensor… | 18.33:1 | 3.0:1 | #000000 | #E8F0FE | Pass |
| h2 Explore the ecosyste… | 18.33:1 | 3.0:1 | #000000 | #E8F0FE | Pass |
| h2 Stay connected | 18.33:1 | 3.0:1 | #000000 | #E8F0FE | Pass |
| h2 Start building with … | 18.33:1 | 3.0:1 | #000000 | #E8F0FE | Pass |
| h3 What's new in Tensor… | 18.33:1 | 3.0:1 | #000000 | #E8F0FE | Pass |
| h3 What's new in Tensor… | 18.33:1 | 3.0:1 | #000000 | #E8F0FE | Pass |
| h3 Introducing Wake Vis… | 18.33:1 | 3.0:1 | #000000 | #E8F0FE | Pass |
| h3 MLSysBook.AI: Princi… | 18.33:1 | 3.0:1 | #000000 | #E8F0FE | Pass |
| h3 What's new in Tensor… | 18.33:1 | 3.0:1 | #000000 | #E8F0FE | Pass |
| h3 What's new in Tensor… | 18.33:1 | 3.0:1 | #000000 | #E8F0FE | Pass |
| h3 TensorFlow.js | 18.33:1 | 3.0:1 | #000000 | #E8F0FE | Pass |
| h3 LiteRT | 18.33:1 | 3.0:1 | #000000 | #E8F0FE | Pass |
| h3 tf.data | 18.33:1 | 3.0:1 | #000000 | #E8F0FE | Pass |
| h3 TFX | 18.33:1 | 3.0:1 | #000000 | #E8F0FE | Pass |
| h3 tf.keras | 18.33:1 | 3.0:1 | #000000 | #E8F0FE | Pass |
| h3 Kaggle Models | 18.33:1 | 3.0:1 | #000000 | #E8F0FE | Pass |
| h3 TensorFlow Datasets | 18.33:1 | 3.0:1 | #000000 | #E8F0FE | Pass |
| h3 TensorBoard | 18.33:1 | 3.0:1 | #000000 | #E8F0FE | Pass |
Methodology: The top 20 text elements by font size were checked. Background color was sampled from the desktop screenshot using a 5-point pattern. WCAG 2.1 AA requires 4.5:1 for normal text and 3:1 for large text.