Katalon True MCP Server

Last updated: October 2026

This document lists tool descriptions of Katalon True MCP server, and explains how to connect to the server using an external AI provider.

Overview

Katalon True integrates with modern AI tools through the Model Context Protocol (MCP).

With the Katalon True MCP Server, you can use your AI assistant (Claude, GitHub Copilot, or ChatGPT with MCP support) to work directly with Katalon True artifacts, such as Test Cases, Test Suites, Test Runs, Test Folders, Requirements, Defects, Iterations (Sprints/Releases), Repositories, and Projects.

This integration lets you create, update, move, duplicate, delete, and search Katalon True assets, as well as query configuration and quality insights, all from within your AI tool—streamlining test management, reporting, and analysis workflows.

Connect to the MCP Server

Availability Notice

• The Katalon True MCP Server integration is available only for Katalon True.
• If you are using TestOps Legacy, please contact your Customer Success Manager (CSM) for upgrade guidance and compatibility support.

Requirements

• Ensure your AI tool supports MCP server connection.
• For tools that do not support remote MCP servers natively, install the required utilities such as npx.
• Confirm that your AI tools (e.g., Claude Desktop, ChatGPT) is configured to connect with an external MCP server.

To Katalon True MCP

To connect your AI tool with Katalon True, point it to the Katalon True MCP endpoint:

• Recommended: use the canonical endpoint https://platform.katalon.io/mcp in any AI tool that supports configuring a remote MCP server natively. This single endpoint works for every workspace: on first connect it starts a browser sign-in and then prompts you to choose your workspace/account (the same picker as logging in to Katalon True), so you do not need to know your subdomain in advance.
• Dedicated domains only: if your organization uses a dedicated Katalon domain, you can substitute https://<your-sub-domain>.katalon.io/mcp instead. See Sub Domain for more information.
• If your tool does not support remote MCP servers natively (for example, Claude Desktop), add the following JSON configuration to your MCP server setup.

Claude Code

Run the following command in your terminal:

claude mcp add katalon-prod-mcp -- npx -y mcp-remote https://platform.katalon.io/mcp --transport http-first

Alternatively, add the server to your project's .mcp.json file:

{
  "mcpServers": {
    "katalon-prod-mcp": {
      "command": "npx",
      "args": [
          "-y",
          "mcp-remote",
          "https://platform.katalon.io/mcp",
          "--transport",
          "http-first"
      ]
    }
  }
}

Codex

Add the following entry to your ~/.codex/config.toml file:

[mcp_servers.katalon-prod-mcp]
command = "npx"
args = ["-y", "mcp-remote", "https://platform.katalon.io/mcp", "--transport", "http-first"]

Claude Desktop

{
  "mcpServers": {
    "katalon-prod-mcp": {
      "command": "npx",
      "args": [
          "-y",
          "mcp-remote",
          "https://platform.katalon.io/mcp",
          "--transport",
          "http-first"
      ]
    }
  }
}

Copilot

{
  "servers": {
    "katalon-prod-mcp": {
      "url": "https://platform.katalon.io/mcp",
      "type": "http"
    }
  },
  "inputs": []
}    

To Katalon Documentation MCP

Use the public MCP endpoint for Katalon Documentation if you need to search or fetch public Katalon documentation from an MCP-compatible client.

{
  "mcpServers": {
    "katalon-docs": {
      "url": "https://mcp.katalon.com/mcp",
      "type": "http"
    }
  }
}

The public Katalon documentation MCP endpoint does not require authentication. To learn more, see Katalon MCP Server.

After updating the configuration:

1. Restart your AI application to detect the new MCP server connection.
2. Your browser will display a prompt asking you to log in to your Katalon True account, then to choose the workspace/account you want to work with.
3. In your AI tool, make sure to toggle on the katalon-prod-mcp tool to activate it.

Supported tools

The Katalon True MCP Server currently supports the following tools:

Workspace detail tools

Use these tools to understand your current workspace context before taking action in a project.

Tool	Description
list_projects	List Katalon True projects the user has access to.
list_repositories	List repositories under a Katalon True project.
find_iterations	Find the iterations (sprints or releases) of a Katalon True project.
find_execution_profiles	List execution profiles for the given projectId and testProjectId (repository ID).
get_user_context	Retrieve the currently authenticated user's profile information.

ALM requirement tools

Use these tools to work with requirements pulled from your integrated ALM tool.

Tool	Description
find_requirements	Find requirements from integrated issue tracking systems (JIRA, Azure ADO).
read_requirement	Retrieve a specific requirement by its internal ID or source system key (e.g. JUL-3449, TO-12345).
link_requirements_to_test_case	Link requirements to a specific test case.
unlink_requirements_from_test_case	Unlink requirements from a specific test case.
find_test_cases_by_requirement	Find test cases linked to a requirement. Supply exactly one identifier: requirement_id (the requirement's internal ID in Katalon True, returned by find_requirements or read_requirement), key (the ALM key, such as JUL-3449), or requirement_external_id (the externalIssueId value). Zero or more than one identifier returns a validation error.
find_alm_integration_projects	List ALM integration projects configured for a Katalon True project.

ALM defect tools

Use these tools to turn failed test outcomes into actionable defects in your integrated ALM tool.

Tool	Description
create_defect	Create a new bug or new defect in the Alm project and link it to a specific test result.

Test artifact management tools

Use these tools to organize and manage testing assets used across your project.

Tool	Description
find_test_cases	Find test cases within a repository.
read_test_case	Get the details of a specific test case by its ID.
create_test_case	Create a manual test case with optional test steps.
move_test_case	Move a test case to a new folder while keeping its existing name.
duplicate_test_case	Duplicate a test case.
delete_test_case	Delete a test case.
update_test_case	Update metadata and content of existing test cases.
find_test_suites	Find test suites within a repository.
read_test_suite	Get details of a specific test suite by its ID, including list of test cases within test suite.
manage_test_suite	Manage test suites and their relationships with test cases.
find_test_suite_collections	Find test suite collections within a repository.
find_test_folders	Find test folders within repository of project.
manage_test_folder	Manage test folders.

Tools that create, move, duplicate, or update a test case, test suite, or folder return result, id, and resource_url. Use resource_url to open the item in Katalon True. When an update changes several test cases, the response holds one link per case.

Automated test run tools

Use these tools to prepare, schedule, and manage automated executions across supported environments.

Tool	Description
build_run_configuration	Build an execution environment configuration (DESKTOP_AND_MOBILE_BROWSER, MOBILE_NATIVE, or API) for create_test_run. Runs locally and creates nothing.
build_schedule	Build a one-time or recurring schedule payload for create_test_run. Runs locally and creates nothing.
create_test_run	Create an automated test run (mode="automated") for a test suite or test suite collection, run now or on a schedule. Pass automated.test_run_type with a canonical value such as DESKTOP_AND_MOBILE_BROWSER, the same value used in build_run_configuration. Unknown or legacy field names are rejected with an error that names the field to send.
read_execution_test_results	Retrieve job-grouped test results for a Katalon True execution.
rerun_execution	Re-run a test execution with optional job targets.
list_schedule	List saved schedules for a project.
update_schedule_configuration	Update an existing schedule's configuration or timing.
read_schedule_detail	Read the current schedule definition (ScheduleResource) for a given id.

Manual test run tools

Use these tools to launch and follow manual test runs, including AI-assisted execution flows.

Tool	Description
read_auts	Return available Application Under Test (AUT) entries with their nested environments.
create_test_run	Create a manual test run (mode="manual") from test suites or test cases, run now or on a schedule. Call read_auts first to choose an AUT environment. A run-now manual run returns the IDs that create_manual_ai_session needs.
create_manual_ai_session	Open an AI-assisted session for an existing manual test run.
read_manual_ai_session	Get the current status of all runner sessions in a manual AI execution.
read_manual_test_run_detail_by_order	Search manual test executions by execution order.

Cloud execution tools

Use these tools to set up and manage Cloud executions.

Tool	Description
list_test_cloud_environments	Fetch environments from Test Execution Cloud API with optional RSQL filter and pagination.
list_test_cloud_agents	List available Test Execution Cloud agents scoped to the current account and project.
list_test_cloud_apps	List uploaded Test Execution Cloud apps with optional RSQL filters and pagination.
manage_test_cloud_dynamic_app	Manage Test Execution Cloud dynamic apps.
delete_test_cloud_app	Delete one or multiple apps from Test Execution Cloud.
get_upload_test_cloud_app_url	Returns a temporary upload URL so you can upload an app file to Test Execution Cloud via curl, then register it for use.
create_test_cloud_app	Create a Test Execution Cloud app from an uploaded app file.
list_test_cloud_tunnels	List active Test Execution Cloud tunnels.
prepare_test_cloud_tunnel_start	Gather download URLs, API keys, and commands to guide users through the Test Execution Cloud tunnel setup experience.
build_private_local_testing_config	Build a private local testing configuration payload.
stop_test_cloud_tunnel	Stop an active Test Execution Cloud tunnel via DELETE /tc/ees/tunnels/{id}.

Test result query tools

Use these tools to investigate execution outcomes and review run history.

Tool	Description
find_test_results	Find test results (execution test results) within a project.
read_test_result	Get the details of a specific execution test result.
list_executions	List test executions with optional filtering.
read_execution	Get detailed information about a specific test execution.
list_execution_outputs	List execution outputs or discover recent executions.

Test result update tools

Use these tools to record manual test results, review automated results, and attach evidence files to test steps or test cases. See the tool schema for supported fields, statuses, and constraints.

Tool	Description
update_test_results	Takes the test run ID from the run URL and one update object.
prepare_artifact_upload	Prepare an evidence file upload and return a signed upload URL. Files can be up to 10 MiB.
confirm_artifact_upload	Confirm that a prepared upload finished and return the artifact ID to use as evidence.

info

update_test_results records a manual result or reviews an automated result; it does not run tests or change test cases. Manual results require an active run assigned to you; automated reviews require a completed run. Upload evidence first with prepare_artifact_upload and confirm_artifact_upload. Treat only applied and unchanged as successful outcomes.

Insights tools

Use these tools to understand insights evaluate quality trends, coverage, and stability across your testing activities.

Tool	Description
get_defect_insights	Retrieves defect insight for a specific sprint/release iteration or time range.
get_requirement_insights	Retrieves requirement insight for a specific sprint/release iteration.
get_test_case_insights	Retrieves test case insight for a specific sprint/release iteration or date range.
get_test_configuration_insights	Retrieves configuration insight for a specific sprint/release iteration or date range.
analyze_test_case_stability	Analyzes stability of test cases identifying flaky tests, consistently failing tests, newly introduced failures, and slow-running tests.
analyze_test_flakiness	Analyze test executions over a lookback period and minimum number of test runs.

TrueTest tools

Use these tools to find captured user sessions, launch test case generation, and work with TrueTest User Journey Maps from your MCP-compatible AI client.

Tool	Description
list_truetest_auts	Find the TrueTest applications under test (AUTs) linked to a project.
list_truetest_tracking_environments	Find the tracking environments available for a TrueTest AUT.
search_truetest_sessions	Search captured user sessions by date, environment, page URL, session attributes, session ID, or number of steps.
get_truetest_session	View the details and replay steps of a selected user session.
create_truetest_test_case_from_session	Submit one to 25 selected captured sessions for TrueTest test case generation. This write operation requires your approval and returns the launch status without waiting for generation to finish.
search_truetest_journey_maps	Search User Journey Maps by environment, status, tracked-data period, creator, or creation date.
get_truetest_journey_map	View a selected User Journey Map, including its pages, flows, test steps, source sessions, and linked test cases.
create_truetest_journey_map	Generate a User Journey Map for a selected environment, time range, and filters.
get_truetest_journey_map_request	Check the progress of a User Journey Map generation request.
create_truetest_test_case_from_flow	Submit one existing Journey Map flow for test case generation, with optional tags and a TC ID prefix choice. This write requires approval and returns the launch status.

You can describe the outcome you want in a prompt. For example:

• Find Staging sessions from August 18 that visited /checkout and contain at least 20 steps.
• Generate TrueTest test cases from these selected Staging sessions.
• Show the active User Journey Map for Staging and summarize its most-used flows.
• Generate a User Journey Map from last week's Staging sessions that visited /checkout.
• Show the checkout flow in my Staging Journey Map, then generate a TrueTest test case from that flow after I approve.

Your AI client may ask you to select a project, AUT, or tracking environment when more than one matches. You can access only the TrueTest data permitted by your Katalon True role.

For flow-based generation, first discover the map and flow with search_truetest_journey_maps and get_truetest_journey_map. Approve the selected flow before submitting it. A STARTED response means generation was queued. Read the flow again with get_truetest_journey_map to inspect its current status and test case link.

Manage TrueTest custom-element configuration

Use these tools to manage element patterns, custom scripts, and action mapping rules for an AUT linked to your selected project. Mapping rules change how matching recorded actions are converted into generated test steps.

Tool	Description
list_truetest_element_patterns	List the AUT's element patterns.
get_truetest_element_pattern	Read a selected pattern and its locators.
create_truetest_element_pattern	Create an element pattern using the locators you provide.
update_truetest_element_pattern	Update approved fields of a selected pattern.
list_truetest_custom_scripts	List the AUT's custom scripts.
get_truetest_custom_script	Read a selected script and its arguments.
create_truetest_custom_script	Create a custom script using the code and details you provide.
update_truetest_custom_script	Update approved fields of a selected script.
list_truetest_mapping_rules	List the AUT's action mapping rules.
get_truetest_mapping_rule	Read a selected rule and its pattern and script references.
create_truetest_mapping_rule	Create a rule for matching actions and an element pattern. A TRANSFORM rule uses a custom script; a REMOVE rule omits matching actions.
update_truetest_mapping_rule	Update approved fields of a selected rule, including its enabled state.

1. Select your project and AUT. Use the client code returned by list_truetest_auts.
2. Ask the agent to list the current resources and read the item you want to change. For example: List my AUT's mapping rules and show the element pattern and script used by the checkout rule.
3. Supply the locator, script, or rule changes you need. Review the proposed changes and approve each create or update operation.
4. Read the saved resource to confirm the change. For a new TRANSFORM rule, create or select an element pattern and custom script from the same AUT before creating the rule.

Skills for AI agents

The MCP server gives your AI agent the tools. The open-source Katalon True skills give it the playbooks: tested, reusable workflows that turn a request such as "cover this requirement" or "ship my Playwright results" into a complete, traceable run on Katalon True.

Supported agents: Claude Code, OpenAI Codex, GitHub Copilot, Cursor, Kiro, Windsurf, Cline, Continue, and other agents that support reusable instruction files.

Included skills

Skill	What it does	Try this prompt
katalon-platform-setup	Installs, verifies, and troubleshoots the MCP connection. Confirms project and repository access, and diagnoses authentication errors.	"Set up Katalon MCP and verify my projects."
katalon-trueplatform-testing	Runs the full requirement-to-result workflow: analyze a requirement, design test cases, import them, build a suite, run with AI, and report results.	"Test requirement CEL-6 end to end on Katalon Platform."
katalon-create-test-cases	Designs ISTQB-aligned manual test cases from a requirement or free-text behavior, checks existing coverage, and imports only what is missing.	"Create manual tests from requirement CEL-6."
katalon-execute-test	Creates manual test runs, starts Run with AI, schedules automated suites, and summarizes pass/fail/blocked outcomes.	"Run a Katalon suite with AI and summarize results."
katalon-test-case-to-playwright-script	Converts manual test cases into Playwright TypeScript automation with Page Object Model, keeping Katalon case IDs for traceability.	"Generate a Playwright script from a Katalon test case."
katalon-playwright-execute	Runs Playwright tests with the Katalon reporter configured and verifies the uploaded run in Katalon True.	"Execute a Playwright suite and upload the report to Katalon Platform."
katalon-upload-report	Uploads Katalon Studio/KRE, JUnit XML, and Playwright reports, then verifies the uploaded runs.	"Upload my JUnit report to Katalon and verify the run."
katalon-release-analyze	Assesses release readiness from coverage, execution health, defect risk, and stability data, with a Ready / Ready with risk / Not ready verdict.	"Is Sprint 24 ready to ship?"

Onboard your agent

Let your agent set itself up. Paste this one line into it:

Fetch and execute the appropriate instructions to set me up for Katalon True from https://docs.katalon.com/agent-setup/prompt.md

The agent fetches the predefined setup instructions, picks the branch for the tool it is running as, installs the skills and the MCP server, and then verifies the connection by listing your real projects. You only need to sign in and, in Claude Code, run /reload-plugins.

This is the same command copied when you click Onboard your agent button in Project Settings > AI.

Install manually (Claude Code)

claude plugin marketplace add katalon-labs/true-skills
claude plugin install katalon-true-platform@katalon-true-platform-marketplace

The plugin installs the skills and connects them to the Katalon MCP server. For install steps for the other agents, see the Katalon Labs repository. Start with katalon-platform-setup to connect and verify the MCP, then use the workflow skills.

Usage guidelines

• Use MCP tools for Katalon Platform data such as projects, requirements, test cases, suites, runs, results, and defects. Use browser or Playwright tools only for the application under test.
• Confirm the target project and repository before creating or updating test artifacts. A repository in Katalon True is the Test Project.
• Understand the boundaries. Requirements are created in Jira or Azure DevOps and then synced to Katalon. Defects must be linked to a failed test result. AI execution can also depend on the application, account, or test environment.
• Include evidence for failures when available, such as the error message, screenshot, or trace.
• Protect credentials. Authentication uses browser-based OAuth. Do not paste passwords, API tokens, cookies, JWTs, MFA codes, or OAuth callback URLs into chat, and do not commit them.

Sample workflow: from requirement to test results

A single prompt such as Test requirement CEL-6 end to end triggers the following flow:

1. The agent verifies the MCP connection and resolves your project and repository.
2. It reads requirement CEL-6 from your integrated ALM tool and analyzes the expected behavior.
3. It designs manual test cases based on the expected behavior.
4. It checks existing coverage and imports only the missing cases, linking each one back to the requirement.
5. It creates or reuses a test suite, creates a manual test run, and starts Run with AI.
6. It monitors the run until every test case completes, then reports results with links: for example, 6 passed, 1 failed, 1 blocked, plus recommended next actions.

To continue into automation, ask the agent to convert the same suite into Playwright scripts and run them. Automated results can be uploaded to Katalon True so they appear next to your manual runs. See Playwright integration for reporter setup details.

What the MCP can and cannot do

The Katalon True MCP Server covers test artifact management, manual and automated execution, results, and quality insights. A few operations are intentionally out of scope:

Not available through MCP	What to do instead
Create requirements	Create them in Jira or Azure DevOps, then sync and link them in Katalon.
Inspect the live application under test	Use your agent's browser or Playwright tooling to explore the application.
Create a defect without a failed test result	Run the test first; defect creation links to a failed test result ID.

Example use cases

Once your MCP server connection is active, you can use your LLM to interact directly with Katalon True. Common use cases include:

• Plan and launch regression runs: ask the assistant to build run configurations, create schedules, and rerun selected executions.
• Investigate release risk faster: ask the assistant to identify flaky tests, newly introduced failures, slow-running tests, and defect trends for a sprint or release.
• Turn failures into action: trace a failed result back to its linked requirement, review related test cases, and create a defect in Jira or Azure DevOps from the same conversation.

These workflows help teams move from simple content lookup to AI-assisted test operations, combining project context, traceability, execution control, and quality analysis in one place.

Watch how Claude Desktop connects to the Katalon True MCP Server to analyze project data.

Troubleshooting

• Invalid JWT token format error on first connection: this is the authentication challenge before the browser sign-in completes, not a server failure. Finish the browser login and retry.
• Connection fails with --transport http: use --transport http-first in the mcp-remote arguments.
• Skills installed but no Katalon tools appear: the MCP server is not connected. Add katalon-prod-mcp manually as shown in Connect to the MCP Server, then restart your AI tool.
• Agent cannot find your project: ask it to run the setup skill ("Set up Katalon MCP and verify my projects") to confirm project and repository access.

See also

• Katalon AI Assistant
• Katalon's built-in MCP servers