Ukrainian engineer Volodymyr Obrizan has created a tool that should reduce the time that test automation engineers spend searching for reasons why autotests fail.
Ukrainian engineer Volodymyr Obrizan has created a tool that should reduce the time that test automation engineers spend searching for reasons why autotests fail.
According to Obrizan, test automation is an «underserved market»: while there are hundreds of tools for a regular backend developer, there are almost none for an automation engineer, tailored to his specific tasks.
He formulated the specific problem as follows: the pytest test runner generates a detailed report on the reasons for test failures only after the entire test suite has been executed — even if the suite has been running for hours. Because of this, the engineer either idles or switches to other tasks, losing context, and only at the end receives poorly structured diagnostic information: large tracebacks, long exception messages, a mixture of test code and library code in one canvas of text.
The solution was to develop an open source plugin pytest-testinel that streams events and diagnostic information to a cloud server as tests run, while simultaneously uploading screenshots, videos, and Playwright traces to the cloud — all of which are available immediately after a specific test completes, rather than after the entire suite completes. Installation does not require refactoring existing tests — it’s a drop-in plugin: registration, pip install pytest-testinel, and one environment variable.
Obrizan solved the second part of the problem — poor diagnostic structure — by breaking the traceback into separate frames, classifying them (test / page object / Selenium, etc.) and hiding the library code that rarely needs to be viewed, as well as pulling screenshots, videos, and traces directly into the report so that the engineer does not have to look for them separately on the CI server.
According to the founder, the team has been using the solution for half a year and has repeatedly recorded cases when a commit with an autotest fix appeared in git before the entire test run had time to complete.
Obrizan admits that the idea of similar tools (ReportPortal, Testomat, TestDino, and others) arose long before him, and in the comments under the publication on DOU, several direct competitors have already been pointed out to him. However, the author intends to develop the product commercially, focusing specifically on diagnosing and correcting errors in web tests — without repeating the functionality of task trackers like TestRail. The plans include developing a skill and a CLI utility for integrating Testinel with AI agents, since, according to Obrizan, the agents themselves have not yet solved the problem of real-time event streaming and cloud storage of diagnostics.
For teams testing with Python, pytest, Selenium, and Playwright, the service is free to use; you can try the product at testinel.dev .
