Platform updates 2026 – Level Access Help Center

Welcome to the Level Access Platform updates!

Note: Additional release notes might be available for specific platform tools and extensions, such as browser extension, Level CI, Level Docs, or mobile functionality. For a full list of additional release notes, refer to What's new?

Version 2026.7.0

Reporting Agent: Iterative Conversation-Driven Report Authoring

Our Reporting Agent now supports multiturn, conversational report editing. You can refine reports through back-and-forth dialogue with the agent instead of regenerating them from scratch. Stay in the platform, keep full control over the narrative, and never lose work. Reports and chat history are saved automatically. Return anytime to continue editing. Stop report generation at any point to redirect the Reporting Agent. For details, refer to Custom reports.

Common findings public API

Prerequisite: Organization must opt-in to Common Findings.
You can now retrieve common findings data programmatically via two new public REST API endpoints. This replaces the need for manual exports or unsupported internal endpoints, giving you a clean, reliable way to pull common findings into your own reporting tools, dashboards, and workflows. You can:

List Common Findings: Retrieve a paginated, sortable, filterable list of Common Findings for your organization
Get Common Finding Details: Retrieve detailed information for a specific Common Finding by ID

API capabilities:

Filter by testing tool (defaults to Access Engine) and digital asset IDs
Pagination with configurable page size (default: 25)
Sorting by severity (default) or other supported fields
Standard API key authentication via x-auth-token

Level Access Design Companion: New annotation workflow and AI-assisted design guidance

Use accessibility annotations to document requirements that might not be evident from the visual design alone. Leverage AI-assisted guidance to support accessibility during the design process.

New accessibility annotation workflows in Design Companion: Designers can now document image alt text, heading order, focus order, and additional accessibility requirements directly in Figma, with annotations displayed on the canvas to support clearer design-to-development handoff.
AI-assisted design guidance: Design Companion can now automatically generate focus order and image alt text annotations. Expanded findings also provide suggested alt text for images and accessible replacement hex values for supported color contrast issues.

The Mobile Engine, On-Device Testing (ODT), and the new ME SDKs (ME SDK iOS, ME SDK Android) now support tablet devices on both iOS and Android. Previously validated only on phone-sized devices, accessibility scans have been extended and verified on tablet viewports, including iPad simulators and Android tablet emulators.

The improvements include:

Mobile Engine rules validated on tablets: The full rule library has been validated against tablet viewports on iOS and Android, with fixes for rules affected by larger screens, different layouts, and tablet-specific interactions.
ODT tablet support: Device discovery, scan sessions, screenshot capture, and metadata push to the Level Access Platform have been validated for iPad simulators, Android tablet emulators, and physical tablet devices.
ME SDKs tablet accuracy: Scan accuracy has been verified against tablet UI trees on both platforms, and sample configurations now include tablet device examples.
Tablet device metadata in LAP: Scan results pushed to the platform now include device type and screen dimension metadata, making it easy to distinguish tablet findings from phone findings.

A11y Testing: Mobile | Mobile Engine ruleset integrated into two new SDKs: ME SDK iOS, ME SDK Android

Launching with Guadalajara, accessibility scans powered by the Mobile Engine ruleset with be available through the new ME SDKs (ME SDK Android, ME SDK iOS), achieving rule parity with On-Device Testing (ODT). Previously, our Level Access SDK relied on Appium and Access Engine rules applied through an XML-to-HTML conversion, which could lose mobile-specific attributes and introduce processing overhead. These new ME SDKs use the same native mobile rules engine available in ODT.

Key changes include:

Rule parity across mobile tools: The same Mobile Engine rule library version running in ODT now runs in the ME SDKs, eliminating inconsistencies between mobile testing surfaces.
Full version traceability: Scan findings include Mobile Engine version, rule version, device metadata, and scan origin, ensuring complete audit traceability.
Platform integration: Findings from ME SDK scans (for both Android and iOS versions) using the Mobile Engine are correctly ingested by the Level Access Platform under the associated digital asset with proper source and metadata tagging.
Updated packages: Updated SDK packages for Android and iOS are published to Cloudsmith and installable via standard package managers, with no breaking changes for existing customers.

Version 2026.6.0

Real-Time activity updates for Ask Level AI Chatbot

Enhanced the Ask Level AI chatbot to provide real-time visibility into its actions as it processes user requests.

Reporting agent: Stakeholder-ready reports (export, share, regenerate, customize)

We've enhanced our Reporting Agent and made it easier for you to export, share, or regenerate AI-powered reports in stakeholder-ready formats, using flexible filtering and comparison options. For example, you can filter findings by compliance standard (for example, WCAG 2.1, Section 508, EN 301 549), set any start and end date (within a 365-day range), and include multiple data levels in your report (digital assets, workspaces, organization). This reduces time from report generation to decision making and remediation. For more details, refer to Custom reports.

VPAT preview: Updated criteria distribution table and document order

The VPAT preview now features a redesigned Criteria Distribution by Standard section and an improved document layout for better readability and compliance tracking.

Note:

Conformance data is now broken out per applicable standard (WCAG → Section 508 → EN 301 549), each with its own conformance score and criteria count table.
Each standard sub-section displays a conformance score (0–100) with delta comparisons against the prior VPAT and the baseline (first) VPAT, so users can track progress over time.
The criteria count table includes four columns — Conformance level, Criteria (#), vs. prior VPAT, and vs. baseline — across Supports, Partially supports, Does not support, and Total assessed rows.
Standards not covered by the evaluation are omitted automatically.

VPAT preview: Auto-applying a specific template for tested standards

An auto-generated VPAT preview now automatically uses a specific template based on the standards included in your manual evaluation. This ensures that your report reflects only what was tested, improving accuracy and credibility for stakeholders. The selected template aligns the VPAT output with the evaluation scope, removing irrelevant standards, sections, and cross-references. As a result, generated reports include only the applicable standards and content.

Note: We support multiple VPAT editions (instead of International-only).

Scan-completed webhook now includes coverage and timing data

The dp_scan_complete webhook event now includes three new fields, enabling fully automated end-to-end accessibility scan workflows without follow-up API calls:

Field	Type	Meaning
pagesScanned	Integer	Total number of distinct pages scanned in the completed scan.
scanStartTime	Timestamp	When the scan began.
scanEndTime	Timestamp	When the scan finished.

These fields are included across all scan types: on-demand scans, monitoring scans, user flow scans, and scans triggered from the browser extension (BE). The webhook is delivered only after the scan has fully completed, ensuring that the payload always reflects final results. Integration systems can now confirm scan coverage, record timing, power reporting dashboards, and run automated gating logic directly from the webhook payload, with no additional API calls required. For more information, refer to Webhooks.

Version 2026.5.0

By default, Level Access Platform uses vertical navigation menu. For more details, refer to Platform navigation and Platform navigation refresh 2026. Until July 2026 you can still use the horizontal navigation system. To switch to the horizontal navigation system, display your profile, and deselect Show the new navigation experience. For details, refer to Edit your profile.

File management at the digital asset level

File management is now tied to individual digital assets. To view or manage files, select a digital asset and, under Manage, select Files. For details, refer to Digital asset files (create, view, download).
Previously, file management was available at the workspace level.

Auto-VPAT uses the correct template for tested standards

Auto-generated VPATs now automatically uses the correct template based on the standards included in your manual evaluation. This ensures that your report reflects only what was tested, improving accuracy and credibility for stakeholders. The selected template aligns the VPAT output with the evaluation scope, removing irrelevant standards, sections, and cross-references. As a result, generated reports include only the applicable standards and content.

Note: We support multiple VPAT editions (instead of International-only).

Platform UI updates

New Banner for API Token Expiration Reminders: A persistent banner in the Generate API Token modal informs Organization Administrators and Developers when API token expiration reminders are sent.
Example banner: Expiration reminders are always sent 15 and 7 days before the token expires.
Visual design refinements: Buttons across multiple pages have been updated to align with brand guidelines and improve contextual awareness.
Navigation area update: The company logo has been removed from the bottom of the vertical navigation menu.

PDF Accessibility Checker updates

Ability to flag and scan corrupted files: The platform now flags structural rule file corruption, if the file fails an ISO 3200 7.5 rue. We’ll scan the file for accessibility, where possible, but the file owner still needs to fix the file. For details, see Fix corrupted files before the upload.

Contrast testing in Test Accelerator

Test Accelerator now includes contrast testing tools to support color contrast validation during evaluations. Contrast testing is available within the evaluation workflow and through the Browser Extension.

AI-powered findings in Test Accelerator

Test Accelerator now includes AI-powered findings writer for web evaluations. You can generate contextual finding text based on the elements selected during an evaluation. After generation, you can accept the suggested text, cancel the result, or restore the original finding text.

Version 2026.4.0

New default navigation (vertical navigation)

We've introduced a new and more intuitive way to navigate the platform! When you log in, a vertical navigation menu appears on the left side of the screen.

The menu is organized into three levels—organization, workspace, and digital asset—and provides quick and contextual access to platform features and pages. Menu options might vary based on your user group, digital asset type, and assigned permissions. Only users with the appropriate permissions can access all the features and menu options.

Note: To switch to horizontal navigation, select Try new navigation in the top-right corner. Select it again to return to the vertical navigation. You can toggle between the new and old navigation systems for a 90‑day period. After that, the horizontal navigation system will no longer be available. For more information, refer to Platform navigation, Platform navigation FAQs and Platform navigation refresh 2026.

Use open-prompts to generate your custom reports

We’ve enhanced the Reporting Agent to give you greater flexibility and control over how you explore and communicate your organization accessibility data. For more details refer to Custom reports.

You can now generate reports that reflect the unique context, needs, and accessibility maturity of your organization.
The Reporting Agent now supports an open-prompt, conversational reporting experience, which allows you to generate AI-powered reports without being limited to pre-selected topics. Use the new Custom Reports section to create, save, mark as favorite, and manage reports for ongoing use.
The existing Create Report → select topics flow remains available, for customers who prefer a guided experience.

Ask Level AI: UI enhancements and accessibility updates

We’ve made several refinements to the Ask Level AI interface to improve visual consistency, usability, and accessibility. We’ve refined typography and layout, icons and controls, accessibility and labeling, and the Ask Level AI input area.

Example improvements:

All icon buttons now include tooltips on hover and ARIA labels (“Close”, “Minimize”, “Expand”).
We now identify agent messages via ARIA labels such as “Level AI response: Welcome, how can I help you today?”

Streamlined task workflow

You can now access your tasks directly within the digital asset level, from the vertical navigation menu. You are no longer redirected to the workspace level. To access the tasks, first select a digital asset. Under Find and Fix, select Tasks. The My tasks page is displayed. Prerequisites: You are logged in, have access to a digital asset, and the vertical navigation is enabled.

On-Device Testing (ODT)

On-Device Testing is now available through Level Access Desktop, enabling you to test mobile applications directly on physical and virtual Android and iOS devices from your macOS environment. This launch introduces guided setup for required development tools, device detection, mobile app configuration, and iOS app resigning for physical devices—making it easier to validate accessibility earlier and more accurately within real mobile experiences.

Level Access mobile engine

Level Access's mobile engine is a native-first testing engine with a comprehensive recommendation library for the following languages and frameworks: Android, Jetpack Compose, iOS, SwiftUI, Flutter, React Native, .NET MAUI (Android and iOS), and Xamarin. For more information, please refer to the Mobile engine testing rules. Please note at the moment this is only available through On-Device Testing.

Outpost

Outpost allows you to run accessibility scans from your own environment instead of relying on externally hosted infrastructure. You can install and manage Outpost within your existing workflows, connect it to your repositories and environments, and send scan results directly to the platform for review. This gives you more control over how scans run while keeping results centralized within the platform.

Confidential Scanning

We added a new Confidential Scanning setting at both the organization and digital asset levels. When the setting is enabled, sensitive data is not saved—screenshots are disabled, and only HTML tags and data attributes are retained. This allows teams to safely scan live sites that contain private or sensitive information.

Version 2026.03

Simplified management of platform

We've simplified user management, which is now available with the Management screen, under the Users tab. Organization Administrators can now easily view and manage different levels of platform hierarchy, and assign or remove users for each level.

Intelligent management of findings across different versions of Access Engine (Single unified finding)

We've eliminated duplicated or related findings, resulting from scans made by different versions of the Access Engine tool. Previously, such orphaned findings remained visible in dashboards and reports, leading to inaccurate data insights. Metrics and dashboards now reflect a more accurate view of your digital asset health. This update affects all areas where common findings are tracked or used as metrics, including Program Dashboard, Organization Report, Dashboard, and Tasks. As a result, you may notice a decrease in total findings, changes in trends, or tasks that display fewer linked findings. These updates are expected and reflect cleaner, more reliable data. For more details, refer to Single unified finding.

New PDF accessibility validation and reporting

We’ve developed a new standards‑based engine that validates PDF documents for accessibility compliance. The engine ensures that all PDFs comply with the PDF/UA1 (ISO 14289-1) standard. As well, we’ve improved the resulting report, specifically the number of rules being checked (106) and the user interface descriptions of those checks. These improvements help you to interpret failures more effectively and identify the steps needed for compliance.

Renamed the option for accessing the product portal

You can access the product feedback portal from the Level Access Platform. Previously, you needed to select the Roadmap option from the platform Help menu. We’ve renamed it to Submit feedback. For details, go to Share feedback on the product roadmap.

Renamed the Edit workspace page to Workspace details

When a user with adequate permissions selects the Edit workspace button, the following page is displayed: Workspace details. Previously, the page was called Edit workspace.
Permission: Organization administrator who can access Organization settings, and then Organization details.

Removed redundant breadcrumb links

We’ve removed unnecessary breadcrumb links from both the Scan results page and the Manual evaluation page.

Access Engine 2.29

Access Engine version 2.29 introduces 19 new rules and updates 18 existing rules to improve automated testing coverage and reduce false positives. These changes strengthen validation across several areas, including:

Autocomplete attributes, such as invalid token values and incorrect token order
ARIA roles and states, including required attribute validation
ARIA attribute usage and restrictions
Accessible names and duplicate ID references in ARIA relationships
Heading hierarchy within modal dialogs

These updates improve standards alignment and increase the reliability of automated findings. As a result, you may notice changes in issue counts and scan scores in future scans. Some scans may identify new issues because of expanded coverage, while others may show fewer issues because of improved detection accuracy.

Version 2026.2.0

Auto-generated VPAT previews

Limited availability. Available only to select customers and for the Accelerate and Enterprise packages. If you have interest in testing this feature, contact your Customer Success Manager.

When the feature is enabled, you can request and download a VPAT preview, generated from the evaluations data available in the platform. A VPAT preview is an unverified preview of your evaluations or manual audit. Upon request, our AI-powered VPAT Agent consolidates the available evaluations data and provides a fully structured, standards-aligned preview report for download. The report is delivered in the MS Word format. VPAT previews are not reviewed or certified by third‑party accessibility specialists and do not include an official compliance seal, unlike expert-evaluated VPATs. For details, go to VPAT previews.

Removed the ability to add a digital asset from the Workspace Portfolio

You can no longer add a digital asset from the Workspace Portfolio page. The Add digital asset button has been removed. Digital assets can be added through the Admin portal by organization administrators.

Common Findings availability update (Essentials and Basic packages)

Previously, Common Findings were available in the Enterprise and Accelerate packages. The feature is now also available:

As a core feature in the Essentials package.
As an add-on feature in the Basic package.

Common Findings unify insights across all your assets to expose systemic issues and accelerate root‑cause remediation.

Multi-value filtering of digital asset findings

We've enhanced filtering of digital asset findings. You can now use multiple identifiers in a single interaction, including Finding ID, Task ID, Page URL, and Finding Title. This enhancement eliminates the need to apply filters one value at a time and repeatedly re-run searches. It also supports bulk inputs, such as values generated by AI-assisted flows, while maintaining predictable and explicit filtering behavior.

For details, refer to Guide: Manage and prioritize automated findings for a digital asset.

Version 2026.1.0

Ask Level AI: Inline side panel and Full-Screen mode

Ask Level AI now supports an embedded side panel and full-screen mode. The change enables multitasking and improves usability without obstructing the platform content. You can easily switch between the reference mode (side panel) and focus mode (full screen that covers the entire viewport for detailed reading or complex queries). The chat sessions are persistent, that is, chat remains active while you navigate across workspaces, dashboards, or assets.

To expand the chat to full screen mode, select the Expand icon. To collapse the chat to reference mode (side panel), select the Sidebar icon. For details about the chatbot, go to Ask Level AI Chatbot.

Exposing Reporting Agent data to automated workflows (through MCP Server)

You can connect the Reporting Agent platform data to external workflows, analytics platforms, and AI models, to enhance flexibility and control over how you use the data. The feature affects data across various platform levels (organization, workspace, and digital asset). The data is exposed through the MCP server. You can access this data from the platform interface, through the Ask Level AI chatbot, or without using the platform user interface. For more details, go to Connect Reporting Agent data to automated workflows through MCP Server.

Known limitation: Version 2026.01 does not support role-based permissions, nor E-charts.

AI prioritization to decide what to fix first

You can now invoke the Prioritization Agent directly from the Digital asset findings page. This enhancement streamlines remediation planning by giving you a quick way to analyze and identify what to fix first. AI prioritization is a decision‑support tool, and not an automated decision maker. Your judgment and oversight remain essential.

To use it, open the Digital asset findings page, select Prioritize with AI, specify a remediation goal, choose a prioritization strategy, and select Apply. The platform then filters and prioritizes the findings to present a focused, relevant subset aligned with your chosen strategy.

Common Findings now generally available in the Accelerate and Enterprise Packages

The Common Findings feature is generally available, for the users of both Accelerate and Enterprise packages. Previously, the feature was available only for the Enterprise package users. Common Findings leverages the power of artificial intelligence (AI) to intelligently group recurring issues across all the digital assets in your entire organization. This intelligent grouping significantly reduces the time it takes to consolidate, sort through, and prioritize findings that affect multiple digital assets by collapsing all your data into a single report.

Ability to configure dynamic field mapping for issue tracking integrations (JIRA, Asana, Azure)

Accessibility tasks pushed to Jira, Azure DevOps, and Asana now include key details—severity, WCAG criteria, and identifiers—in dedicated fields. This eliminates manual cleanup and ensures tickets are ready for filtering, workflows, and automation. Admins can configure supported fields to receive specific task information. Tickets are created with structured, query- ready data in place. The content of the ticket Description field remains unchanged for backward compatibility. For more details, go to Configure field mapping for issue tracking integrations.

Known limitations: Only text-based fields are supported in this release (no other field types). The current implementation of dynamic mapping behavior is supported and validated for tasks originating from scans and monitors findings. Tasks created from manual findings and design evaluation findings cannot be dynamically mapped, because manual findings data is stored in a different collection and contains a different set of fields.

Support for custom headers for Jira Data Center OAuth 2.0 connections

We now support custom headers for Jira Data Center connections, enabling compatibility with network-level authentication systems.
Context: Some Jira DC instances require service-token headers on every incoming request. You can now configure custom header name/value pairs when configuring the OAuth 2.0 connection for Jira Data Center. Header values will be stored securely and applied to all subsequent outbound Jira requests. For more details, go to Set up a Jira connection.

Ability to reauthorize existing Jira Data Center OAuth 2.0 connections

For Jira Data Center (v8.22+) with OAuth 2.0, organization administrators can now:

Reauthorize existing connections using stored or new Client ID and Secret.
Remove Jira DC OAuth 2.0 or Jira Cloud connections in Unauthorized or Interrupted state.
Complete setup for Jira DC instances that require additional network-level authentication headers.

Previously, credential issues (rotation, expiration, compromise, or denied consent) left connections unrecoverable and instance URLs locked.

Hardware Test Guide available in Test Accelerator

Manual hardware accessibility testing is now fully supported within Test Accelerator. Testers can select from a vetted hardware library, follow guided testing steps, capture evidence, and log findings without leaving the platform. The Hardware Test Guide supports both Section 508 and EN 301 549, improves consistency across evaluations, and removes the need to rely on AMP for hardware testing workflows.

On this page: