StudioAssist Ask mode
This document explains how to use StudioAssist Ask mode in Katalon Studio.
- Katalon Studio Enterprise version 10.1.0 onwards.
- An AI provider selected. For more details, see StudioAssist Preferences.
StudioAssist Ask mode enables instant interaction with Katalon Studioβs AI assistant right within your workspace. Use it to:
- Learn Katalon Studio features
- Write test scripts
- Learn built-in keywords
- Reinforce testing concepts
- Explain a code snippet
- Ask follow-up questions
Supported AI modelsβ
To choose your preferred AI model, click the dropdown in the Model field via the KSE configuration window, when applicable.
If your account has no configured AI model, the following notification appears: There is no AI model in your account. You can then click Request Access to contact your account owner to enable AI features, or add your own API key.
Starting version 10.3.0, we have added more AI models. See StudioAssist preferences for more information.
For OpenAI compatibility information, refer to OpenAI Supported Models.
Get started - Use Ask modeβ
-
Open StudioAssist window.
StudioAssist chat window (Ask Mode) automatically opens on startup, on the right panel:
If StudioAssist chat window is not visible or if it's closed, click the StudioAssist icon on the main toolbar to open.
- Prompt and see results.
Common use casesβ
- Ask Testing Questions: Get instant answers to your queries about automation testing concepts. For example:
How to create a test case in Katalon Studio?
- Generate Test Scripts: Specify clearly your application type, test objects, and steps.
Write me a web test script:
1. Open browser to URL: GlobalVariable.G_SiteURL
2. Click the Make Appointment button
3. Enter Username and Password
4. Click Login button
5. Close browser
- Explain Code Snippets: Request concise explanations.
Summarize this code in 3 sentences.
// Paste your code here
- Create Custom Keywords: Provide detailed requirements.
Write a custom keyword for login handling:
- Method name: login
- Accept TestObjects for username, password fields, login button
- Accept Strings for username, password
- Explain a custom keyword: If your team uses custom keywords in test scripts, StudioAssist can help you quickly understand what each keyword does.
Explain what this custom keyword does:
// Paste your custom keyword code here
- Reuse custom keywords: If you've already generated custom keywords, you can reference them directly in your request to build new test scripts.
Use the clickElement custom keyword to tap on the login button.
- Troubleshoot issues: Clearly mention errors and steps.
What should I do if my test case fails at step 3 due to timeout?
- Write manual test cases: Include descriptive information.
Provide a manual test case outlining the steps to test a basic login process following a password reset. The test case should include a descriptive title, preconditions, test steps with expected outcomes, pass/fail criteria, and any relevant notes for handling errors or exceptions.
- Optimize code: Clarify optimization intent.
Please optimize this code for better performance:
for (int i = 0; i < array.length; i++) {
if (array[i] % 2 == 0) {
System.out.println(array[i]);
}
}
(Mobile testing) Write a mobile test scriptβ
If you have captured mobile objects using Mobile Object Spy and want to write test cases for basic scenarios, you can provide the app ID and object list in a manual script.
Write me a mobile test case with the following steps:
1. Start application, uninstalling the application automatically after run
2. Tab on the linear layout, timeout 0
3. Get text 'HorizontalNestedScrollView', timeout 0
4. Close the app
My app ID: 6261b88f-275b-4ae5-b0d1-b95fbf2560bd
Object list:
findTestObject('Object Repository/android.widget.LinearLayout')
findTestObject('Object Repository/android.widget.TextView - HorizontalNestedScrollView')
(API testing) Verify status codeβ
To generate an API test case, ensure you specify the API endpoint (for example, https://reqres.in/api/users?page=2) and the corresponding test object (getAllUsers) explicitly in your request.
Write me an API test case with the following steps:
1. Send a GET request to the URL 'https://reqres.in/api/users?page=2'.
2. Verify that the response status code is 200.
Use the object ID: findTestObject('getAllUsers').
(Web testing) Generate a Login test script using existing test objectsβ
If you already have test objects prepared in your project, you can quickly generate a web test script using StudioAssist. Simply provide the necessary test object details, variables, and the steps you want to automate.
For an example of predefined test objects, refer to our sample project: Sample WebUI project (Healthcare sample).
I have a list of test objects as below:
- Make appointment button: 'Page_CuraHomepage/btn_MakeAppointment'
- Username: 'Page_Login/txt_UserName'
- Password: 'Page_Login/txt_Password'
- Login button: 'Page_Login/btn_Login'
- Appointment div element: 'Page_CuraAppointment/div_Appointment'
I also have a URL: GlobalVariable.G_SiteURL, and two variables `Username` and `Password`
Write me a test case to perform the following steps:
1. Open browser to the URL stored in G_SiteURL
2. Click the make appointment button
3. Fill in the username and password fields based on the variables
4. Click the login button
5. Close the browser
Attach files and imagesβ
You can now include files from your project or local machine directly in a conversation to give StudioAssist more context.
-
In the StudioAssist chat window, click the paperclip icon to attach files.
-
Choose one of the following:
- Select from project: Browse test cases, test suites, listeners, Groovy/Java files, feature files, etc.
- Select from computer: Upload external files such as CSV data files, images and screenshots.
- You can send up to 10 files per question.
- Each file must be text-based and under 10MB.
- Invalid files are flagged and won't be sent.
- Attachments appear above the input field and are also shown under each sent message.
- After you sent the file(s), StudioAssist will return the result according to your request.
Use current file as contextβ
In addition to manual file attachment, StudioAssist automatically includes the content of the file you're actively working on as context. This removes the need to manually copy-paste file content into the chat.
In the chat window, the Current file box is selected by default to include the file you're working on. When you switch between files, the context also automatically updates. If you want to turn this off, simply untick the box.
This feature only applies for the following files:
- test case
- test suite
- dynamic test suite
- test listener
- web service request
- custom keyword
- BDD feature file
- profile
- step definition (
.groovyfile) - Java class (
.javafile).
Tips for getting better responsesβ
To get the best results when asking StudioAssist to generate a test case, provide clear and specific details. Here are some tips:
- Clearly specify your input in certain format. See Tips to use StudioAssist.
- Clearly specify the type of application under test (Web. Mobile, API): Here are some example prompts:
Write me a mobile test case.Generate an API test case. - IInclude details about the AUT: Mention specifics like the app ID for a mobile application or the URL for a web application.
- Enable Auto-include project context information in StudioAssist Preferences to auto-populate object IDs.
- Regularly clear conversations when context changes. All files attached during a conversation are used as context until the conversation is cleared. The more files included, the higher the chance that StudioAssist may generate less relevant results.
- Set optimum complexion tokens to 16000 for optimal responses.
Known limitationsβ
While StudioAssist chat window offers valuable support, Ask mode currently has a few limitations:
- Limited knowledge base: StudioAssist does not yet pull information directly from Katalon Studioβs official documentation. For complex or uncommon queries, use Agent mode.
- Project context limitations: StudioAssist is not aware of global variables, custom keywords, other test cases, test data, etc. To make sure that the code is generated with proper variables and locators, we recommend specifying them in your question, or use Agent mode.
- Single conversation mode: StudioAssist supports only one active conversation at a time. Refresh the chat window (close/re-open) as you switch contexts for better accuracy.
- AI response accuracy disclaimer: Responses may be inaccurate or misleading due to hallucination, a common issue in LLM AI models. Always verify AI-generated output before implementation. You may rate responses as Good or Not Relevant to help improve StudioAssist. However, your feedback is not used to train the AI model or track your conversations.
- Performance: In projects with large number of test objects (more than 200 objects), StudioAssist tries to enumerate every object, which can cause timeouts and failed retrieval attempts.