Skip to main content

Testiny MCP Server

The Testiny MCP (Model Context Protocol) Server enables AI tools like Claude (Code, Desktop), VS Code with Copilot, and other MCP-compatible applications to interact directly with your Testiny project data. This allows AI assistants to create and read test cases, create test runs, update test results, and manage your testing workflow seamlessly.

Watch a short overview of the Testiny MCP Server setup and in action:

Why use Testiny with AI assistants?​

The Testiny MCP Server lets AI assistants work directly with your test management data. Instead of manually searching for test cases, creating runs, or updating results, you can describe what you want in natural language and let your AI assistant perform the work.

Connect Testiny to Claude, ChatGPT, GitHub Copilot, Cursor, or any MCP-compatible client to:

  • Create and organize test cases
  • Search and analyze existing test repositories
  • Generate new tests from requirements or user stories
  • Create and manage test runs
  • Execute tests and update results
  • Track testing progress across projects

Setup​

To use the Testiny MCP server, you'll need to configure it with your Testiny credentials and add it to your AI tool's MCP configuration.

Prerequisites​

  • A Testiny account with API access
  • A Testiny paid or trial plan
  • An optional API key from your Testiny API Keys settings
  • If no API key is used, the Authorization header is not needed and OAuth will be used instead
  • MCP Server URL: 
    https://app.testiny.io/api/v1/mcp-server
    For Testiny Server, please use your instance URL instead of app.testiny.io.

Configuration Examples​

Here are examples of how to configure the MCP server with specific clients. You can use the MCP server with any MCP compatible client (protocol '2025-06-18').

Configure the Testiny MCP server in your VS Code settings mcp.json or workspace configuration:

{
  "servers": {
    "testiny": {
      "url": "https://app.testiny.io/api/v1/mcp-server",
      "type": "http",
      "headers": {
        "Authorization": "Bearer your-api-key-here"
      }
    }
  }
}
Security Notice

Treat your API key like a password and keep it secure. The API key inherits the same permissions as the user who created it. Consider creating dedicated API keys with limited project access for MCP server usage.

Available Tools​

The Testiny MCP server provides the following tools for AI assistants to interact with your Testiny data:

Project Management​

  • listProjects - Retrieve all accessible Testiny projects
  • listTestCaseFolders - Browse test case folder structure within projects
  • listTestCases - Get test cases with optional keyword filtering
  • listTestRuns - Get test runs with optional keyword filtering

Test Cases & Folders​

  • createFolder - Create a folder with a title, and optionally a description and parent folder
  • editFolder - Edit the title or description of a folder
  • moveFolder - Move a folder into other folders
  • createTestCase - Create test cases (in a folder) with a precondition, custom fields and content in the Steps, Text or BDD template
  • editTestCase - Edit the title, precondition, template, content and custom fields
  • moveTestCases - Move test cases into (other) folders

Test Execution​

  • createTestRun - Create a test run for organizing test execution
  • setTestResult - Update a test case result (passed, failed, blocked, skipped) in a test run
  • setTestResultBatch - Update test case results (passed, failed, blocked, skipped) in a test run
  • getTestStatus - Get the result, step results, comments and assignee of one or all test cases within a test run
  • addTestResultComment - Add a comment to a test result

Troubleshooting​

If you encounter issues with the MCP server:

  1. Verify API Key: When using an API key, ensure your API key is valid and has appropriate permissions
  2. Ensure OAuth enabled: Ensure that OAuth is enabled in your Testiny organization settings. If your organization restricts the allowed redirect URLs, also enable the redirect URLs of desktop apps for clients such as Claude Desktop, Cursor or VS Code — see External App Access / OAuth2
  3. MCP feature available: Make sure you have a paid Testiny plan or are in the trial phase
  4. Check Base URL: Confirm the base URL matches your Testiny instance (Cloud vs Server)
  5. Project Access: Verify the API key's user has access to the projects you're trying to access

For additional support or feature requests, please contact our support team as explained here.

Use Cases​

Test Cases as Context​

Your test repository contains valuable knowledge about how your application works. Test cases often capture business rules, edge cases, validation requirements, and expected behavior that may not exist in specifications or documentation.

With the Testiny MCP Server, AI assistants can use this information as context when helping with development and testing tasks.

For example:

"Generate API tests for the new subscription feature based on our existing billing test cases."

An AI assistant can:

  1. Search related billing and subscription test cases
  2. Identify existing patterns and business rules
  3. Generate consistent new test cases
  4. Create them directly in Testiny

Faster Test Planning​

When new requirements arrive, AI assistants can help identify what needs testing. You can also connect with other MCP servers (for example, Jira MCP or GitLab MCP) to leverage project information, requirements, issues and merge requests.

Example prompt:

"Review these requirements and suggest missing test cases based on our current regression suite."

The assistant can compare requirements against existing test coverage and propose additional test cases or exploratory testing ideas.

Streamlined Test Execution​

AI assistants can also help manage execution activities.

Example prompt:

"Create a regression test run for Sprint 24 and mark all authentication tests as passed."

The assistant can locate the appropriate project, create the test run, and update results through the MCP tools.

Example Workflows​

Generate Test Cases from Requirements​

When new functionality is introduced, AI assistants can transform requirements into structured test cases and create the test cases directly in Testiny.

"Create test cases for password reset functionality in the Web Application project."

"Review the requirements TNY-123, TNY-124 and TNY-125 and create test cases in project Web Application."

Prompt Hint

Use specific prompts, e.g. define the Testiny project you're working in.

Analyze Test Coverage & Investigate Failed Tests​

Understanding what is tested, what might be missing, and where failures occur is critical for maintaining quality. AI assistants can analyze your test repository and execution results to provide quick insights.

"Show me all test cases related to checkout and identify potential coverage gaps."

"Show me all failed tests from test run 'Release v1.25.0' from project 'Demo' and summarize the affected areas."

Augment a Release Test Run with Release Changes​

Before executing a release test run, AI assistants can cross-reference the release's changes — release notes, merge requests, commits since the last release tag — against the test cases in the run.
Combined with a source-code MCP server (for example, GitLab MCP), the assistant adds a focused "what to verify" comment to each affected test case and highlights changes that are not covered by any test case in the run.

"Augment the test run 'Release v1.42.0' in project 'Web Application': check the code changes since the last release tag, add a comment to each affected test case with what to verify, and report changes that are not covered."

The AI assistant will:

  1. Find the project and test run using listProjects and listTestRuns
  2. Read the test cases in the run using getTestStatus and listTestCases
  3. Retrieve and summarize the release changes through the connected source-code MCP server
  4. Add a verification-focus comment per affected test case using addTestResultComment
  5. Report changes that have no matching test case in the run
Reusable Skill

Recurring workflows like this are a great fit for reusable instruction files. See our example skill file and adapt it to your needs.

Assisted Test Automation Integration​

You can ask your AI assistant to help automate test result reporting:

"Mark all smoke tests in today's regression run as passed and add a summary comment."

or

"Create a new test run called 'Sprint 24 Regression Tests' in the Mobile App project and execute the test case TC-123"

The AI assistant will:

  1. Find the Mobile App project using listProjects
  2. Create the test run using createTestRun
  3. Try to execute the steps in the specified test case
  4. Update the test result using setTestResult
  5. Add comments using addTestResultComment
Prompt Hint

Always include enough context information for the assistant to be able to use e.g. a browser MCP server with URLs and test logins.

By combining Testiny's test management capabilities with AI assistants, teams can spend less time on repetitive administration and more time improving test quality.

Code Generation Example​

Generate test automation code based on your Testiny test cases:

"Generate Selenium WebDriver test code for test case TC-456 in Java."

The AI assistant will:

  1. Retrieve the test case details using listTestCases
  2. Analyze the test steps and expected results
  3. Generate appropriate automation code
  4. Suggest how to integrate with your test reporting using setTestResult
Prompt Hint

Include information about specific libraries, code style and existing test structure like page objects in your prompt.

Good to know​

Test Case Templates​

AI assistants can create and edit test cases in all three test case templates:

TemplateContent fieldContent
StepsstepsA list of steps, each with an action and an expected result
Texttext.stepsText, text.expectedResultTextTwo free-form documents: one for the steps, one for the expected result
BDDbddScenariosGherkin scenarios as plain text, as in a .feature file
  • createTestCase uses the template set in template. If no template is set, it uses the template that matches the content provided, or Steps if no content is provided.
  • listTestCases returns the template of each test case and the content of that template under the same field names.

BDD scenarios are stored as sent: the MCP server does not validate the Gherkin. A Feature: line is optional, and a # language: directive sets the Gherkin dialect. Without one, the dialect is detected from the keywords.

Rich-Text Formats​

Descriptions, preconditions, comments, step cells and Text template documents are rich-text fields. Every tool that reads or writes them accepts a format option:

  • markdown (default) — simple to read and write, but loses some formatting such as underline and text color
  • xml — keeps all the formatting, so an edit doesn't lose any of it

The MCP server provides a format reference for each format as an MCP resource (tny://format/markdown and tny://format/xml). AI assistants read these references before writing, and use xml to read and write a field that already has content.

Avoiding Conflicting Edits​

Test cases and folders returned by listTestCases and listTestCaseFolders include a conflict tag (_etag). The create, edit and move tools also return the new tag.

If an AI assistant passes this tag to editTestCase or editFolder, the edit only goes through if nobody has changed the test case or folder since the assistant read it. If someone has, nothing is written, and the tool returns the status update_conflict and the current tag, so the assistant can read the item again first. Without a tag, the edit overwrites the stored content.

Last updated on