March , 2026

🚀 New Features added

🤖 Agentic Mode

State the goal in plain language and Qase figures out the path, breaking down natural language into actionable steps, captures screenshots for visual feedback, and generates test code.

High-level goals: "Verify a user can purchase shoes on our marketplace" and the rest is handled for you.
Precise targeting: Specific instructions like "Click 'Get a Demo' and fill the email field" are handled directly

📁 Test Advisor in Folder View

Test Advisor is now available in Folder view with bulk-run support. Select all test cases from a folder and analyze them in one go.

Green: Ready to automate with minimal oversight
Yellow: Good potential but missing detail
Red: Lacks the steps or specificity automation requires

🔁 Parameterized Test Conversion

Convert data-driven manual tests into parameterized automation automatically. One test scenario with multiple input combinations becomes one reusable flow.

Automatic detection: Qase identifies the parameterized structure from multiple input combinations
Data extraction: Extracts values and generates one reusable test flow
Execution instances: Creates individual runs for each combination automatically

🔌 API Testing

Qase can now execute GET and POST API requests via CURL directly in your test flow, enabling end-to-end validation of both UI behaviour and the underlying API calls in a single test.

CURL support: GET and POST requests run inline within the test flow
Random Data Generator: Creates unique emails, usernames, and passwords on the fly to keep environments clean

🔧 Self-Correcting Generation

When your application changes mid-generation and a step breaks, Qase AI no longer stops and hands the problem back to you, instead:

Failure detection: Automatically detects when a step breaks mid-generation
Auto-recovery: Analyses what changed and re-attempts with a corrected approach
Continuous generation: The process keeps moving without requiring manual intervention

🎨 Canvas Support

Heavy canvas elements — tools like Miro or Figma — have always been automation dead zones. Qase AI now interprets visual elements within a canvas just like standard DOM elements.

Visual element interpretation: Canvas elements are treated the same as standard DOM
Coverage gap closed: Products with canvas-based UIs can now be fully included in automation

September, 2025

🚀 Export test code in Python

You can now export your recorded tests to Python and it uses the new GPT-5 AI model for better results. The tests naming is handled properly with spaces and support for special characters.

Python + Playwright: Full support for generating Python Playwright tests with proper async/await patterns

Python + Selenium: Complete Python Selenium test generation with WebDriver integration
XPath Support: Python Selenium exports now use XPath selectors for better element targeting

📈 Everything should be just a pinch faster now

1. Faster Action Resolution:

Parallel Processing of multiple action candidates simultaneously.
Smart Triangulation: Better algorithm for determining click points and scroll actions
Modular request handling reduces processing time

2. Streaming Reliability:

Streaming timeouts now adapt based on AI response patterns.
Improved Error Handling: Better recovery from network interruptions during AI streaming

3. Smarter AI Integration

Model Optimization: Improved model selection based on task complexity
Enhanced Prompts: More specific prompts reduce AI ambiguity and improve response quality

🐛 Bug Fixes

1. Form Filling Improvements

Forms now properly clear existing content before filling new values
Better handling of rich text editors and content editable elements
Improved behavior with pre-filled forms and empty fields.

2. Validation & Error Prevention

Fixed crashes when elaboration_required field is null in chat processing
Clearer restrictions on cross-domain navigation to prevent test failures
Better validation for URL opening scenarios in intention validation

3. Data Generation and Job Execution

Fixed randomizer to properly consider existing placeholders
Action executor now properly includes target host information to fix redirection issues

🚀 New Features added

1. Get notified when your AI-generated tests are ready

2. Smart Randomized Data

The generator now intelligently detects when to use dynamic data instead of hardcoded values.

Based on context, automatically identifies fields that should use random data (emails, usernames, etc.)
Automatically remembers and reuses generated values across test steps for consistency
Added exhaustive support for different random data patterns for emails, names, numbers, dates and a lot more.
Enhanced placeholder replacement with validation to prevent test failures