Kernector helps QA engineers turn a User story into reviewed test cases.
Writing that coverage by hand is slow, and the cases drift as the issue changes. Pasting the issue into a generic chatbot is faster, but the output is ungrounded and has no review step.
The engineer sends a GitHub Issue URL or a Jira key, keeps or drops the suggested coverage, and chooses manual or Cucumber for each test. Kernector reads the issue live, writes the cases only after that selection, and stops without writing if the issue has changed. The cases can then be exported as a file or created as Xray tickets. The same workflow runs in the Next.js app and over MCP in Cursor or Claude Code.
The demo is local. It needs server-side model credentials and a tracker connection, so there is no public app URL. Setup and the walkthrough are in the repository: https://github.com/mahmoudazaid/Kernector (docs/capstone/demo.md).
