Mehmet Engin Sacan is a QA Lead and Senior Test Engineer based in Rotterdam with 10+ years of experience in test engineering, test automation, and QA leadership in Agile teams. He leads quality work end-to-end: defining test strategy, aligning stakeholders, planning risk-based testing, and ensuring releases are stable and predictable. He designs and builds automation frameworks for API, UI, and mobile testing, and connects them to CI/CD pipelines to improve speed, coverage, and confidence. Mehmet is hands-on with Java and Python (also JavaScript/TypeScript) and uses tools such as Selenium, Cucumber, Playwright, Cypress, Robot Framework, RestAssured, KarateDSL, Jenkins, GitLab CI, Docker, and Kubernetes. As a leader, Mehmet focuses on developing people and building a learning culture. He has coached colleagues, delivered internal trainings, supported OKR implementation and recognition practices, and created learning programs for new joiners. He is also active outside of day-to-day work through conference talks and workshops, including sessions on BDD and AI usage in test automation, and he holds two patents (one related to NLP).
Speech: Using MCP server for desktop applications via Pytest Framework
Have you ever wished your AI assistant could run your pytest tests directly? With the Model Context Protocol (MCP), this is now possible. In this article, I’ll show you how to create an MCP server that connects PyTest to AI tools like Claude, making desktop application testing much easier.
What is MCP and Why Should You Care?
MCP is a new protocol that allows AI assistants to interact with external tools and services. Think of it as a bridge between your AI assistant and your development tools. Instead of copying test results back and forth, your AI can run tests, analyze results, and suggest fixes automatically.
Why PyTest + MCP for Desktop Applications?
Desktop applications are complex. They have user interfaces, file systems, and system integrations. Testing them requires:
- Running multiple test scenarios
- Checking UI components
- Validating file operations
- Reviewing logs and screenshots
An MCP server for PyTest helps by:
- Running tests on demand – Ask your AI to run specific tests
- Analysing failures – Get immediate insights into what went wrong
- Managing test fixtures – List and organize your test files
- Reviewing results – See test output formatted clearly
