`.
+ - *What it misses:* It cannot tell if text *looks* like a heading (e.g. large, bold) but is coded as ``. That requires visual analysis.
+
+- **1.4.3 Contrast (Minimum) (Partial)**
+ - *What it catches:* Solid text on a solid background (e.g. computed hex colors).
+ - *What it misses:* Text over gradients, background images, or semi-transparent overlays, where the effective contrast cannot be derived from a single background value.
+
+- **4.1.2 Name, Role, Value (Partial)**
+ - *What it catches:* Invalid ARIA combinations (e.g. `role="button"` without appropriate focusability).
+ - *What it misses:* Whether a custom widget *actually behaves* like the role it claims (interaction logic); that requires manual or assistive-technology testing.
+
+### 4.3 Why some criteria are "manual only"
+
+Criteria that are **not covered** (or only partially covered) by automation typically fall into three groups that are difficult for tools to fully test:
+
+- **Category A: Time and change (dynamic)**
+ Examples: 2.1.1 Keyboard, 2.4.3 Focus Order, 2.4.7 Focus Visible.
+ *Reason:* They require navigating through time and sequence. A static snapshot cannot show whether tab order is logical or whether focus is trapped in a modal.
+
+- **Category B: Human cognition (subjective)**
+ Examples: 1.3.2 Meaningful Sequence, 3.3.3 Error Suggestion, 3.3.4 Error Prevention.
+ *Reason:* Deciding if an error message is "helpful" or if a sequence is "meaningful" requires human judgment about logic and clarity, not just syntax.
+
+- **Category C: Complex visual or physical behavior**
+ Examples: 1.4.10 Reflow, 1.4.12 Text Spacing, 2.5.1 Pointer Gestures.
+ *Reason:* Testing reflow requires rendering at specific viewport sizes (e.g. 320px) and checking for horizontal scroll or overlap; text spacing requires applying user styles and checking for clipping. These need a rendering or interaction environment, not just code parsing.
+
+### 4.4 WCAG version specifics
+
+- **4.1.1 Parsing:** Correctly treated as **obsolete** in WCAG 2.2 (removed from the standard). The rule set reflects this.
+- **2.5.8 Target Size (Minimum):** Often supported partially (e.g. a rule that checks approximate target size); measuring 24×24 CSS pixels is tractable.
+- **2.5.7 Dragging Movements:** Typically **not covered** by automation; simulating drag gestures programmatically is complex. Manual testing or assistive-technology testing is required.
+
+**Summary of depth:** The tool covers **syntax** (HTML/ARIA validation) and uses AI to extend into **static semantics** (e.g. descriptive alt, descriptive title, language match, images of text). It does not replace **dynamic behavior** testing (keyboard, focus, timing) or **complex visual rendering** checks (reflow, spacing); use the **What we do NOT cover** document and manual testing for those.
+
+---
+
+## 5. How to use the platform to test (Web)
+
+Use the TestMu AI Accessibility platform to run automated accessibility scans on your **website or web app**. Follow these steps to test covered items:
+
+| Step | How to use the platform to test |
+|------|-------------------------|
+| **1. Access the platform** | Log in to TestMU / LambdaTest and open the **Accessibility** (or **TestMU AI Accessibility**) section for web testing. |
+| **2. Select URL or page** | Enter the **URL** of the page or web app to test, or select a page from your project. You can run a single-page scan or configure a flow (e.g. multi-page or logged-in state). |
+| **3. Run an accessibility scan** | Start an **accessibility scan** for the selected page(s). The platform runs automated checks against the rules listed in Section 3 (WCAG 2.0, 2.1, 2.2 Level A, AA, and selected AAA as supported). |
+| **4. View the report** | After the scan completes, open the **accessibility report**. Review **violations** (issues found), **passes** (rules that passed), and **incomplete** (items that need manual verification). |
+| **5. Filter by WCAG level or rule** | Use report filters to view issues by **WCAG level** (A, AA, AAA), **rule ID**, or **impact** (Critical, Serious, Moderate, Minor). Use this to prioritize fixes (e.g. Level A and AA first). |
+| **6. Export or share results** | Export the report (e.g. PDF, CSV) or share the report link with your team. Use results to fix issues (see Section 6 for remediation) and re-run scans to verify. |
+
+**Tip:** Combine automated scans with **manual testing** (keyboard, screen reader, zoom, resize) and the **What we do NOT cover** document for full WCAG coverage and compliance.
+
+---
+
+## 6. How to check and how to do (by principle)
+
+For each principle (Perceivable, Operable, Understandable, Robust), below is a concise **how to check** and **how to do** (remediation) for the areas TestMU AI / LambdaTest Accessibility Web covers.
+
+---
+
+### 6.1 Perceivable
+
+**What we cover (WCAG):** 1.1.1, 1.3.1, 1.4.1, 1.4.2, 1.4.3 (2.0); 1.3.4, 1.4.10, 1.4.11, 1.4.12 (2.1).
+
+**How to check**
+
+- Run a full-page (or partial) accessibility scan with TestMU AI / LambdaTest Accessibility and review violations for images, headings, lists, tables, color, contrast, and layout.
+- Manually: Turn on a screen reader (e.g. NVDA, VoiceOver); check that images have meaningful or empty alt; headings and structure are announced; form fields have labels; color is not the only indicator.
+- Check contrast with a contrast checker (e.g. 4.5:1 for normal text, 3:1 for large text and UI components); check at 320px width for reflow and at 200% zoom for resize text.
+
+**How to do (remediation)**
+
+- **1.1.1 (Images):** Add short `alt` text for informative images; use `alt=""` or `role="presentation"` for decorative images. For complex images, provide a short alt plus long description (e.g. in text or `longdesc`/link).
+- **1.3.1 (Structure):** Use semantic HTML (``, `
.default})
3. Install the ACCELQ Agent Dashboard.
-## Integrating LambdaTest From ACCELQ Agent Dashboard
+## Integrating .default})
-4. Now enter your LambdaTest username and access key and save/restart the ACCELQ Agent.
+4. Now enter your .default})
-3. Once the authentication is successful, you can select the real devices that is fetched from the LambdaTest platform.
+3. Once the authentication is successful, you can select the real devices that is fetched from the