Breadcrumbs

Getting Started with Leapwork AI Studio

 Welcome to the Leapwork AI Studio Tech Preview. 

This preview gives a select group of users early access to AI Studio before general availability. You’ve been invited because your feedback, experience, and perspective can help shape the future of the product.

AI Studio is still in active development. Some functionality may change, improve, break, or be removed as we learn from usage and feedback. That is the point of the preview: to learn quickly, improve the experience, and understand what matters most to real users.

During the Tech Preview, we are especially interested in:

  • Which AI Studio capabilities you use most

  • What workflows you try first

  • Where the product helps you move faster

  • Where the experience feels confusing or incomplete

  • Where the AI output feels useful, surprising, incorrect, or untrustworthy

  • What would make AI Studio valuable enough to use regularly

Please join our Slack community to provide feedback and engage with other AI Studio users! We’d love to hear from you. 

What is AI Studio

Leapwork AI Studio is the newest component of Leapwork’s Continuous Validation Platform: an agentic QA and testing solution for creating, running, and managing automated tests for web applications.

AI Studio is designed to help teams move from slow, manual, and brittle testing toward a more continuous model of validation. Instead of starting from a blank canvas, users can generate test cases from natural language, record real user actions, use requirements or documentation as context, co-create automation with AI, edit generated tests, and bring existing Playwright scripts into a shared workspace.

AI Studio is built for modern QA professionals, SDETs, developers, DevOps teams, and technical testers who want to use AI to accelerate test creation and validation while keeping control over the result. It helps teams plan, build, maintain, and execute tests with agentic AI assistance.

Core Functionality and Demos

The spotlights below showcase AI Studio’s primary functionality, and are designed to help you get familiar with the product quickly. For each function, we’ve included a demo application, a sample test case, and a walkthrough video.

If this is your first time using AI Studio, we recommend watching the videos in order before trying your first workflow. You can access the demo applications here: Leapwork Demo Sites.

Recorder

The Recorder functionality is a great place to start if you want to see how AI Studio turns real user actions into automated tests. 

With AI Studio, recording happens directly in the browser, so there’s no local software to install before getting started. You can create an environment, launch a built-in browser, record a workflow, and run the resulting test case from within the tool.

This demo also shows how AI Studio can help maintain and improve tests over time, including self-healing functionality and natural language prompting for creating and running tests. 

Watch the demo below to get familiar with the Recorder workflow before trying it yourself.

In this video, you’ll see: 

  • How to record a test case using the built-in browser  

  • How to run a recorded test  

  • How self-healing helps when a test needs to adapt  

  • How natural language prompts can be used to create and run tests  

Knowledge Base & Blueprint Generator

The Knowledge Base & Blueprint Planner functionality shows how AI Studio uses your existing product context to understand your application and help generate relevant test coverage. 

By connecting or uploading sources such as requirements, documentation, code repositories, ticketing systems, or internal knowledge base tools like Confluence, AI Studio can create a knowledge base specific to your application. It then uses that knowledge base to map functionality and dependencies, answer questions about intended behavior, and generate a test suite with evidence linked back to the original source material. 

Watch the demo below to see how AI Studio turns application knowledge into a test blueprint. 

In this video, you’ll see how AI Studio can: 

  • Build a knowledge base from application documentation and connected sources  

  • Visualize functionality and dependencies  

  • Answer questions about intended application behavior  

  • Generate test coverage based on known requirements  

  • Link test recommendations back to supporting evidence  

AI Studio MCP and AI Agents

The MCP & AI Agents functionality shows how AI Studio can connect directly into modern AI coding workflows, allowing users to create and manage tests from tools like Claude Code, Codex, Antigravity, and other MCP-compatible development environments. 

With AI Studio’s MCP functionality, you can use AI agents to create test suites from prompts, interact with AI Studio outside of the main UI, and bring testing closer to where development work is already happening. Many of the actions available in the AI Studio interface can also be performed through MCP, helping reduce context switching between coding, testing, and validation. 

AI Studio also supports agent-based interactions within the product itself, making it easier to navigate, understand, and work with your test suite through chat. 

Watch the demo below to see how AI Studio brings testing directly into AI-assisted development workflows. 

In the video below, you’ll see how AI Studio can: 

  • Connect with AI coding tools through MCP  

  • Create test suites directly from prompts  

  • Support test creation within development workflows  

  • Use AI agents to interact with and navigate your test suite  

  • Reduce the feedback loop between development and QA  

Glossary 

Below is a glossary to help you get acclimated to AI Studio’s terminology and nomenclature. 

Term 

Definition 

AI Agent 

An AI-powered assistant that helps create, manage, or interact with tests. 

AI Studio 

Leapwork’s AI-native product for creating, running, and maintainingautomated tests for web applications. 

Blueprint 

A test case outline that describes what a test should validate, including high-level steps and expected results. A blueprint can be turned into a single executable test case.

Blueprint Generator 

AI Studio functionality that uses a Knowledge Base to generate multiple blueprints that can be converted into real test cases.

Cloud Browser 

A zero-install, zero-maintenance browser environment managed by Leapwork, used to record and run tests in AI Studio. Cloud browsers can be selected by browser type and geographic region.

Cloud Execution 

Running tests in AI Studio’s cloud environment instead of on local infrastructure. 

Continuous Validation 

Leapwork’s approach to continuously validating that software works as expected as it changes. 

Evidence 

Source material that supports a generated recommendation, Blueprint, or response. 

GA 

General Availability; the broader release stage after Tech Preview. 

Human-in-the-Loop Approval 

A review step where the user approves AI-generated actions before they are executed or adopted. 

Knowledge Base 

AI Studio’s application-specific understanding generated from documentation, requirements, code, tickets, and test assets. 

Knowledge Graph 

A visual map of the functionality and dependencies AI Studio identifies from your Knowledge Base. 

MCP 

A protocol that connects AI Studio to AI coding tools such as Claude Code, Codex, or Antigravity. 

Playwright 

Playwright is an open-source end-to-end testing framework developed by Microsoft, designed for automating modern web applications across Chromium, WebKit, and Firefox browsers.

Prompt 

A natural language instruction given to AI Studio. 

Recorder 

Functionality that captures real user actions in the browser and turns them into an automated test. 

Self-Healing 

AI Studio’s ability to help tests adapt when an application changes. 

Test Case 

A set of steps used to validate a specific behavior or workflow. 

Test Suite 

A collection of related test cases. 

Traceability 

The link between a generated test or recommendation and the source material behind it. 

XPath Healing 

A form of self-healing where AI Studio helps recover or adapt a test when an XPath-based locator changes or breaks. 

Community Guidelines 

Please help us maintain a safe space for the community to engage and collaborate positively by following these guidelines: 

  • Keep preview information private. Do not post screenshots, recordings, product details, or commentary publicly unless Leapwork gives explicit approval. 

  • Share feedback openly and constructively. Bugs, rough edges, missing features, confusing moments, and “this didn’t work how I expected” comments are exactly what we want. 

  • Share what you’re trying, learning, loving, questioning, or breaking so this feels like an active community shaping AI Studio together 

  • Do not share sensitive customer, company, or personal data in Slack.

  • Use the feedback form for product feedback and bug reports so our team can track, triage, and follow up. 

  • Be respectful. This is a builder community. Assume good intent, help others where you can, and keep discussion focused on learning and improving the product. 

  • Discussion of other tools and competitors is welcome when it is constructive, respectful, and focused on learning not harassment, vendor-bashing, or unproductive criticism. 

  • Remember that AI Studio is in Tech Preview. Features may change, break, improve, or be removed as we learn from usage and feedback. 

FAQs 

How do I access the product? 

  • AI Studio Tech Preview access is currently invite-only. If you received an invitation, use the access link below to sign in and begin exploring. 

  • Access AI Studio: https://leapwork.ai 

How do I give feedback? 

Your feedback is one of the most important parts of the Tech Preview.  We are looking for honest, specific, and constructive feedback about your experience with AI Studio. We want to understand what you tried, what worked, what did not work, what felt valuable, and what would make the product better. 

Use the feedback form for product feedback, feature requests, confusing experiences, AI output quality issues, and workflow observations. 

Join our community and post in the #feedback-and-ideas channel: Slack Community 

How do I report a bug? 

If something appears broken, unexpected, or impossible to complete, please report it as a bug. A good bug report helps us understand what happened, reproduce the issue, and prioritize the fix. 

Join our community and post in the #bugs channel: Slack Community 

How do I request support? 

If you’re having trouble accessing the application or something appears to not be working as intended, please let us know and we will be able to support you ASAP. 

Join our community and post in the #help-and-access channel: Slack Community