Skip to content
You're viewing the v2 docs. Looking for v1?Go to v1 docs
Docs
Testsigma
Popular questions
↑↓ to navigate↵ to selectesc to close
Book a demo

Product updates

A full rundown of our latest releases, improvements and fixes in Testsigma

August 2026
Released August 28, 2026Latest

Redesigned test case results page


New features
Error State Actions
When a step fails, the right pane gives one-click access to approve or reject an auto-healed element, compare runs, run visual analysis, or view the root cause, without leaving the page.
Root Cause Analysis
Each failed step has a Root cause block in the Analysis tab, next to Locators, Logs, Step Settings, and Metadata. Testsigma reads the error, screenshot, page source, and any heal attempts, and explains the failure in plain language. Available on accounts with Analyzer V2 enabled.
Compare Runs
Compare the failed step with the same step in any previous run, or the whole test case with a prior run, and compare screenshots and source code across runs. Comparison includes the steps around the failure, and run history labels each run with its status and run type. The current step's comparison appears on the first screen; for any other step, click Compare Runs.
Visual Analysis
Overlay the screenshot captured while authoring on the one captured during execution, so differences such as a shifted button or an open modal stand out.
Improvements
Auto-Heal Review
Approve or reject a healed element on the first screen. Use Update element to fix a locator yourself, or Relearn step to capture it again. When a heal attempt fails, the root cause explains why, and a locator trace shows how each heal was reached.
Resizable Panes
Drag the divider between the step list and the detail pane to resize either side.
Released August 21, 2026

Salesforce support in the Modern engine


New features
Salesforce on the Modern Engine
Salesforce applications are now supported in the Modern engine, alongside Web and Unified Mobile, with Salesforce NLPs and add-ons. To move an existing Salesforce application from the Classic engine, contact your Customer Success Manager.
Released August 20, 2026

Execution videos for local executions


New features
Execution Videos for Local Executions
Local executions now record execution videos, not only screenshots. Review failed runs visually and share them with your team, as with cloud executions.
Released August 19, 2026

Reorganized feature settings


Improvements
Grouped Feature Settings
Related features are grouped into a single setting. For example, Accessibility Testing now covers CSV, VPAT, web, and mobile.
Settings > Preferences
Several features, including AI features, are now managed from Settings > Preferences.
Experiments Tab
Features in beta are listed on the Experiments tab.
Generative AI for All Accounts
Generative AI features are enabled for all accounts and use credits.
Upgrade Icon
An Upgrade icon marks features and preferences that are not included in your plan.
Released August 10, 2026

Export test reports with the REST API


New features
Export Test Reports via REST API
Export test reports in XML, PDF, or Excel at any result level: test plan run, machine, test suite, or test case.
Asynchronous Generation
Send the export request, then poll the same request until status changes from IN_PROGRESS to SUCCESS, and download the report from the returned URL. The download link expires quickly, so follow it right away. A failed generation returns HTTP 200 with "status": "FAILED", so check the status field rather than the HTTP code.
Improvements
PDF Screenshot Options
Choose which screenshots to include (all, failed, or none) and the resolution (low or high). PDF export is available on Enterprise plans; contact support to enable it.
Test Data in Reports
For data-driven test cases, reports show the test data profile and test data set name for each iteration.
Names as of Run Time
Reports and the Run Results page show each test plan, suite, test case, environment, machine, and data profile name as it was when the run happened, so renaming or deleting an item no longer changes older reports. User names always show their current value. Older runs show the current name, and a hyphen (-) marks deleted items.
Deprecations
JUnit Report Endpoint
/reports/junit/<RUN_ID> is deprecated and replaced by /reports/execution_result/<RUN_ID>, which returns XML in the Testsigma schema, not JUnit XML. Old-endpoint responses carry Deprecation, Sunset, and Link headers. If your pipeline parses JUnit output, such as the Azure DevOps Publish Test Results task, keep using the JUnit endpoint until you update that step; its output is unchanged until the sunset date.
July 2026
Released July 20, 2026

Playwright execution and Unified Mobile reach production


New features
Playwright Execution Engine
Web and mobile tests now run on Playwright as the execution engine, executing test steps roughly 60% faster than Selenium.
Unified Mobile
Author one test case and run it on both Android and iOS, instead of maintaining a duplicate suite per platform.
Improvements
Shared Element Repository for Unified Mobile
A single element entry stores both the Android and the iOS locator, so the two platforms cannot drift apart.
Step-Level Platform Control
Turn an individual step off for one platform without splitting the test case in two.
June 2026
Released June 16, 2026

Smarter execution, deeper AI, and enterprise-grade control


New features
Copilot: Run from Any Step
Start execution from any step, with no need to re-run the full setup prefix when debugging a late-failing step in a long test.
Copilot: Pause & Resume
Pause a live run instantly without pre-set breakpoints; the button dynamically switches between Resume and Pause.
Edge & IE Mode Support
Copilot now runs inside Microsoft Edge and Edge IE Mode, so enterprise teams with Edge-restricted apps no longer need workarounds.
Atto: Conversational Task Control
Create, edit, and remove test tasks via natural language prompts before and after browser sessions.
Atto: Confluence Integration
Atto connects to your Confluence workspace to pull documentation and requirements as context for test generation.
Auto-Heal: Shadow DOM & iFrames
Auto-Healing now detects and heals elements inside Shadow DOM and iFrame-embedded content.
Mobile: Copilot & Atto Live Editor
Full AI-assisted execution and Atto Live Editor extended to iOS and Android, with complete feature parity with the web editor.
New Viewer Role
Read-only Viewer role for stakeholders and auditors — full visibility without any edit access.
Improvements
Analyzer Agent in Live Editor
Real-time failure insights directly in the editor; traces upstream steps and highlights the most probable root cause.
Copy, Paste & Clone Steps
Copy and clone individual or multiple steps across test cases; clone directly below or at any position.
WCAG Accessibility Coverage +30%
30%+ of the latest WCAG success criteria now covered across Web, iOS, and Android.
XRay: All Issue Types
All XRay issue types (default and custom) now supported for AI-generated test context, not just "Test".
TestRail: Multiple IDs per Test Case
Link multiple TestRail IDs to a single Testsigma test case; many-to-one mapping was previously limited to one ID.
Persistent Session History & Audit Log
Append-only history log per Atto learning session tracks every task completion, pause, resume, and human response for full auditability.
May 2026
Released May 21, 2026

Debug where the failure is, and see what execution is doing


New features
Execute from Anywhere
Start a run from any step without executing the steps before it, so iterating on step 300 of a 500-step test no longer means re-running the 299-step prefix.
Auto-Healing Simulation
A wavy underline marks the element being healed as it happens, making auto-healing observable during a run rather than silent.
Improvements
Animated Status Banners
Sticky banners report the current execution state and update as it changes, so a paused, healed, restarted, or finished run reads at a glance without opening logs.
Add Step Below
The control now sits on the step itself rather than behind its three-dot menu, which matters most when inserting several steps during a fix pass.
Sync-Aware Recording
Inline recording adds steps in bulk and keeps the execution point in step with the application, so you can record mid-debug and resume without restarting. Recording above or below the execution point raises a re-sync prompt instead of surfacing the drift as a later failure.
Released May 19, 2026

Microsoft Edge support in Copilot and Agentic Learning


New features
Edge in Copilot
Run Copilot sessions in Microsoft Edge or Microsoft Edge (IE Mode). Select the browser when you launch Copilot.
Edge in Agentic Learning
Agentic Learning now supports Microsoft Edge. Edge (IE Mode) is not supported.
Browser Detection
Testsigma detects the browser only when it is installed in the default applications folder.
April 2026
Released April 27, 2026

Mock servers for testing before the backend is ready


New features
API Mock Servers
Create mock endpoints, define mock rules, and run tests against simulated backend responses, across dry runs, plan execution, and local execution.
Improvements
Shift-Left UI Testing
UI tests can start as soon as the frontend is ready, without waiting for backend APIs to be deployed.
Wider Scenario Coverage from One Test
Swapping mock rules lets a single test case validate success paths, error states, timeouts, and edge cases.
Released April 22, 2026

Confluence as a context source, and a calmer Atto session


New features
Confluence as an Atto Context Source
Connect Confluence alongside Figma, Jira, and the test case library as a source Atto generates from. Requires the Confluence integration to be set up in Settings.
Pause and Resume in Copilot
Pause a Copilot run at the current step, act on the application yourself, then resume, without setting a breakpoint first.
Improvements
Atto Asks Only When It Needs To
The "Need Attention" state now triggers only where Atto genuinely needs input, evaluated per session, instead of being raised by default.
No Redundant Library Reads
"Read existing test case library" switches itself off after its first use, avoiding repeated clicks and wasted generation calls.
Generation Timing Is Visible
Each message reports how long its generation took, so response latency can be measured rather than estimated.
Auto-Heal Accuracy
Locator keys that a model invents but the DOM does not contain are now tracked rather than passed through silently, and heal prompts can be updated in production without a deployment.
Released April 20, 2026

On-premises release 8.9.8


New features
Copilot
Live Editor is now Copilot on self-hosted installations: faster launch, automatic arm64 or x64 detection, proxy, SSL, and JVM settings in the UI, real-time execution visibility, and a debug toolbar.
Atto
Higher-quality test generation, updates to existing test cases, validation before saving, real-time debugging, and Auto Heal that learns from history.
AI Capabilities
SOAP API support, Excel (XLSX) validation, Figma frame search, and Salesforce field enrichment.
Accessibility Testing
WCAG 2.2 validation for Web, Android, and iOS, including contrast, structure, and target size checks.
Improvements
Integrations
GitLab bug creation, improved Jira search and fields, Azure DevOps project filtering, and Zephyr custom field support.
Mobile
Flutter support, iOS deep links and keychain control, better WebView interactions, and more stable recording.
Recording
Redesigned Live Editor, Shadow DOM and iFrame capture, reusable step groups, and smart element naming.
Maintenance
Auto Heal for iFrames and Salesforce, AI Analyzer suggestions, and adaptive interaction fallbacks.
API Testing and Security
Real-time API validation, nested schema support, AES-256 encryption, and secure runtime variables.
Authoring
API-based step creation, editing executed steps, reorder and insert support, and encrypted variables.
Reporting
Faster PDF exports, iteration-level details, real-time status, and audit trails.
Reliability
CI/CD stability fixes, serialization and rendering improvements, and better Android connectivity.
Released April 18, 2026

Analyzer decides whether healing is the right fix


New features
Analyzer-Guided Auto-Healing
The AI Analyzer now performs root cause analysis on a failure before healing it, and updates a locator only when a genuine UI change is detected, rather than healing blind.
Non-Native Mobile Browsers
Mobile web test cases can run on browsers such as Chrome or Firefox on Android, not only the device's default browser.
Fixes
Test Plan Errors
Fixed the intermittent "Uh-oh" error raised while changing a test plan.
Go to Project
Fixed the Go to Project button loading indefinitely.
Execution Videos
Fixed execution videos being unavailable for some runs.
Released April 16, 2026

Live emulator recording, custom dashboards, and a read-only role


New features
Mobile Recorder with Live Devices
The mobile recorder now streams a live Android emulator or iOS simulator inside Testsigma, so taps, swipes, and text input are captured instantly instead of through a screenshot-and-refresh cycle. Released as a beta.
Custom Dashboards
Build a dashboard from 30+ widgets covering execution health, reliability, team productivity, AI contribution, and infrastructure performance, each with a chart type suited to its data. Dashboards are shareable and download as a PDF under 25 MB.
Read-Only User Role
View-only access across the product, for stakeholders, auditors, and external teams.
Test Data Profile Folders
Organize test data profiles in a folder hierarchy rather than a flat list.
Improvements
Inspect While Recording
Inspect elements, read the source tree, and debug locators inside the recorder session, with network and debug logs inline.
Edit Steps Inside the Recorder
Reorder, nest, group, and bulk-edit steps without leaving the recording session.
iOS Locator Strategy
iOS recording uses Accessibility ID as its primary locator, falling back to auto-heal where an ID is missing.
Copilot Executor Pause
Pause an active run without terminating the session, to inspect application state as it stands.
Released April 1, 2026

Atto learns mobile and Salesforce, and Terminal maintains itself


New features
Agentic Learning for Mobile
Atto now learns and refines Android and iOS test cases the way it already did for web, refining generated steps against the live application. Mobile authoring reaches parity with web.
Atto for Salesforce
Atto generates Salesforce test cases, creating tests for Flows and Workflows with no prompt needed and taking a prompt for everything else. It reads live Salesforce metadata, so generated tests match the org's real state.
Terminal System Tray Controls
Open Terminal, check for updates, change settings, and toggle startup from the tray, with the signed-in account always visible. Proxy and agent settings can be reset from Configure Settings instead of by deleting the Testsigma folder.
Improvements
Immediate Update Detection
Terminal detects a new version in near real time and prompts to update, rather than waiting for a scheduled check up to 30 minutes away. Mandatory updates install on the next launch.
Build Type Detected Automatically
Terminal downloads the right build for the machine, arm64 or x64, with no manual selection, and reports which one it is running on the Info screen.
Launch on Startup
Terminal can start with the machine.
Relevant NLP Suggestions Only
The step field no longer loads the full action list on click; suggestions appear once what you type matches something, which removes the lag when creating a step.
Auto-Heal for Salesforce Flows
Auto-heal keeps Salesforce Flow tests working as the underlying flows change.
March 2026
Released March 19, 2026

WCAG accessibility testing across web, Android, and iOS


New features
Accessibility Testing
Run WCAG compliance checks from inside a Testsigma test case, on web, Android, and iOS, with no third-party tooling.
Documented Coverage
Web covers 30 of 86 success criteria across all four WCAG principles; Android and iOS each cover 18 of 56, with platform-specific rule implementations. Each rule traces to the criterion it validates.
Released March 5, 2026

Generate, insert, and execute steps inside one Copilot session


New features
Copilot Generator
Generate, insert, and execute steps inline without leaving Copilot. Steps land where they belong in the flow rather than at the end, so there is nothing to reorder afterwards. Web only at this release.
Global Debug Toolbar
A persistent toolbar with Resume and Restart that works on every step, so a paused run can be resumed from anywhere.
Improvements
Add and Execute Individually
Run a single generated step or a block of them without ending the authoring session, keeping the generate, execute, verify, and adjust loop in one place.
Edit Any Step While Paused
Change a step that has already executed, then resume without losing progress, which matters most in large test cases.
Loop Changes Apply Immediately
A modification inside a loop takes effect on the following iterations with no restart.
Execution Stability
Fixes to latency, intermittent failures, and unpredictable execution states.
Released March 2, 2026

Agentic AI Home, cross-project moves, and audit-ready logs


New features
Agentic AI Home
A unified dashboard for Gen AI and autonomous test automation flows.
QTest Integration
Flexible, customizable integration for test management and result syncing.
Cross-Project Import
Move test cases and assets between projects of the same application type.
Export JSON
Export all project artifacts in a single JSON for audit, compliance, or migration.
SSO Login
SSO authentication for secure access.
Improvements
Recorder Experience
Onboarding infographics, scroll support, and a refactored Web Recorder for performance and stability.
Step-Level Accessibility
In-house accessibility checks at the step level for improved accuracy.
Visual Testing Upgrades
Thresholds, region selection, and AI-powered root-cause analysis to reduce false negatives.
Immutable Audit Logs
Exportable, searchable audit logs for compliance and traceability.
Salesforce Flow Generation
Automated test case generation for Salesforce flows and workflows.
February 2026
Released February 25, 2026

Testsigma Terminal for local execution


New features
Testsigma Terminal
A desktop application for local test execution, required to run Copilot for Web, Salesforce, and Atto.
Copilot Sessions
Terminal connects Copilot to your local machine.
Status Dashboard
See whether Copilot is ready and what is running.
Active Executions
Track local dry runs in real time.
Live Logs
View logs as they stream, without collecting log files manually.
Improvements
Automatic Updates
Packages update in the background each time Terminal relaunches.
Configuration
Set proxy, SSL delegation, JVM arguments, and memory allocation from Settings.
Error Recovery
Port conflicts, stuck processes, failed downloads, and broken upgrades are fixed automatically or offered as a one-click fix.
Released February 23, 2026

Atto test case lifecycle upgrades


New features
Run Before Saving
Execute Atto-generated test cases in Copilot before saving them. Validate the run, debug failures, fix locators, and edit steps, then save only test cases that are ready.
Background Sessions
Atto sessions run in the background and stream progress in real time. Leave the page, work elsewhere, and return to see progress.
Improvements
Read Existing Test Cases
Choose whether Atto reads your existing test cases before generating new ones. Turn it on for generation that follows your repository's patterns, or off for faster generation from scratch.
Overwrite Test Cases
When Atto updates a test case, overwrite the original instead of creating a duplicate. The change history is kept, and you can still save as a new test case.
Released February 19, 2026

Auto Heal for Salesforce


New features
Auto Heal for Salesforce
Self-healing is now available for Web, Mobile, Mobile Web, and Salesforce. When a locator breaks on one of Salesforce's dynamic elements, Testsigma sends the DOM, screenshot, and broken locator to an LLM, which finds and applies a working replacement.
Released February 9, 2026

Mid-run editing, Terminal 2.2.4, and more


New features
Edit Executed Steps Mid-Run
Edit, add, remove, or reorder steps that have already run, and continue from the right point.
SOAP API Tests from Definitions
Generate SOAP API tests from service definition files.
GitLab Bug Reporting
Create GitLab issues directly from failed test cases.
iOS Deep Links on App Launch
Launch an iOS app directly to a specific screen with a deep link.
iOS Keychain Control on Upload
Turn iOS Keychain support on or off when you upload an app.
Launch Terminal from the Web
If Terminal is not running when you launch Copilot, start it from the browser. Requires Terminal 2.2.2 or later.
Agent Settings in Terminal
Configure SSL verification delegation, trust store type, proxy, and JVM settings from the Settings screen instead of command-line arguments.
Log Redaction
Sensitive values are masked in logs, while diagnostic details are kept.
Improvements
Live Editor Is Now Copilot
Live Editor is renamed to Copilot across the product and documentation. Functionality is unchanged.
Choice When Closing Copilot
Open results in a new tab, or stay in the original tab, which updates with the results. You can remember your choice.
Simpler Copilot Interface
The refresh option is removed, stopping a session behaves consistently, and Copilot and the recorder offer the same capabilities.
Shadow DOM and iFrame Capture
The recorder reliably captures elements inside shadow DOMs and iFrames.
Local Execution Status in Terminal
Terminal shows the status of executions that use a local agent, and the dashboard refreshes when the run finishes or stops.
Agent Status for Mobile and API
For mobile executions in Copilot, an indicator shows whether the local agent is ready. API runs that use a local agent appear under Active Executions.
MySQL 8.4.x
The MySQL connector supports MySQL 8.4.x.
On-Demand Component Updates
Atto and Windows automation components update without a full reinstall.
Faster Agent Shutdown
The agent shuts down faster when you exit Terminal.
Complete Test Plan Exports
Exported test plan reports include results from every execution iteration.
January 2026
Released January 30, 2026

Atto 2.0 for mobile


New features
Atto 2.0 for Mobile
Atto 2.0 test case generation is available for iOS, Android, and Mobile Web, with all the improvements from the web launch. Generate test cases by providing context, and run them right away.
AI Auto Heal on Mobile
AI Auto Heal adapts to UI elements during execution.
Released January 12, 2026

Redesigned debugger


New features
Global Debug Toolbar
Skip, Resume, and Restart are available for the whole run, not hidden on each step.
Edit Executed Steps
Edit, delete, or add steps after they run, including failed steps.
Loop Debugging
Select a loop iteration from a dropdown to see its step results. Steps added or edited inside a loop run in the next iteration.
Improvements
Resume or Restart
Resume from the next step or restart the test case. New and edited steps run in the next execution.
Smarter Resume
If you delete the last executed step, the debugger works out where to resume. Reordering steps by drag and drop resets their status and uses the new order.
December 2025
Released December 4, 2025

Auto Heal confidence, step group reuse, and Azure AD OIDC


New features
Azure AD OIDC
Sign in with Microsoft Azure AD OpenID Connect.
Step Groups Across Projects
Share and link step groups across projects.
Wait Until Element Not Present
A new wait action for elements that disappear.
Video for Failed Steps
The execution video for a failed step is embedded in the step details.
OAuth 2.0 Grant Type
Support for an additional OAuth 2.0 grant type.
Improvements
GenAI Auto Heal Confidence Threshold
AI-driven locator changes are applied only at 50% confidence or higher, which reduces flaky tests.
Auto Heal Reporting
Reports break down every heal action: Saved, Backup, and AI.
Parallel Execution Stability
Only one local agent can run on a machine at a time.
API and Swagger
Updated APIs and Swagger documentation.
Test Run Sync
Deleted test cases are removed from linked test runs.
WCAG Improvements
Platform updates aligned with WCAG guidelines.
Fixes
Long Test Cases
Fixed an issue when adding steps after the 23rd step.
Test Management Connectivity
Fixed an issue that affected the connection between Testsigma and connected test management tools.
iOS Taps
Fixed tap actions on some iOS tab elements.
November 2025
Released November 25, 2025

Atto 2.0


New features
Better Test Case Generation
Atto generates more accurate test cases. Generation can take a little longer. No action is needed to get the update.
Updates to Existing Test Cases
Atto reviews existing test cases and suggests changes based on new context. Available from December 1, 2025.
Improvements
Improved Analyzer
Every failed step has an Analyzer agent with more useful suggestions for fixing the issue.
Simpler UI
A cleaner, easier-to-use interface.
October 2025
Released October 30, 2025

A new Testsigma Terminal


New features
New Testsigma Terminal
A new version of Testsigma Terminal is available. To upgrade, uninstall the old Terminal, then download and install the latest version from Testsigma. The old Terminal keeps working for now but no longer receives updates or fixes.
Improvements
Faster Launches
Live Editor opens reliably every time.
Agent Management
Terminal recovers from crashes automatically and adapts to browser updates.
Reliability
Sessions stay stable.
Released October 29, 2025

Smart Executions, Flutter support, and more


New features
Smart Executions
Test runs are optimized based on previous results and code changes.
Flutter Support
Support for the Appium Flutter integration driver, with UI support for Flutter elements.
WebView in Native Apps
Switch to a WebView and use web locators, such as XPaths, in native iOS and Android executions.
Encrypted Runtime Variables
Runtime variables are encrypted during execution.
Bluetooth NLPs
New NLPs turn Bluetooth on or off in mobile apps.
Improvements
Zephyr Integration
Improved Zephyr integration.
Back Navigation
Back navigation returns to the previous screen.
Android ADB
More reliable ADB connections.
Azure DevOps Video Links
S3 video URLs in Azure DevOps work items are clickable.
PDF Export and Reports
Smoother PDF export, faster report generation, and accurate test case counts.
NLP Handling
Better handling of text fields and logical operators.
Test Data and Variables
Consistent parameter names and better error handling.
Fixes
Recorder and Debugger
Fixed recorder and debugger crashes, and step editing and action capture on web and mobile.
Accessibility and PDF Reports
Fixed accessibility element highlighting and PDF report performance.
Agents and Execution Videos
Fixed agent connectivity, session creation, and execution video loading.
Integrations
Fixed TestRail add-ons, Jenkins environment selection, and GitLab icon visibility.
Cloning and Importing
Fixed cloning and importing of test steps and elements.
September 2025
Released September 22, 2025

CI/CD reruns, build numbers in emails, and more


Improvements
CI/CD Reruns
Reruns of runs triggered from Jenkins or Azure DevOps use the original CI/CD environment.
Build Numbers in Emails
Test plan emails show the build number and attach the execution report.
.app Uploads
Upload .app files directly to the iOS simulator.
Sorted Runtime Variables
Runtime variables are listed alphabetically.
Azure DevOps Bug Filing
Search and filter projects when you file a bug in Azure DevOps.

Was this page helpful?

Last updated: