Summary

The Cloud and Data Center split is written into the tool surface, so which edition you run decides which of these calls exist for you.

delete_test_case and delete_test_run are marked Data Center; update_test_run, add_test_cases_to_run and list_executions_by_cycle are marked Cloud; create_folder, get_test_execution and search_test_runs behave differently on each. Labelling it per tool is the right call, and it does mean a workflow written against one edition will reach for calls the opposite edition does not have — worth knowing too that on Cloud, search_test_runs returns cycles sitting in the exact folder rather than descending into sub-folders. Authentication spans Zephyr and Jira alike: the Zephyr key and the Jira username and token all have to be present before anything runs.

What it is

A Zephyr Scale server against a Jira-authenticated Zephyr instance, exposing test case authoring, folder management, test run and cycle handling, execution lookup and search, with the differences between Cloud and Data Center called out per tool.

What you get

Test case authoring: create_test_case writes STEP_BY_STEP, PLAIN_TEXT or BDD content and accepts project_key, folder, objective, precondition, priority, status, labels, owner_id, estimated_time, component_id, issue_links, parameters, custom_fields and test_script; get_test_case reads one back by test_case_key; update_test_case_bdd replaces BDD content and preserves fields you leave out.
A way to match your project's own field structure: create_test_case points at the zephyr://testcase/EXISTING-KEY resource so a real test case can be fetched and its shape copied, custom_fields included.
Folder handling: create_folder takes a path such as "/Parent/Child", resolving parent segments on Cloud and sending the full path on Data Center; get_folders returns the entire subtree under folder_path at every depth, or every folder matching project_key and folder_type.
Runs and cycles: create_test_run takes test_case_keys, environment, owner, planned_start_date, planned_end_date, issue_key, test_plan_key and jira_project_version; get_test_run and get_test_run_cases read one back; update_test_run sets owner, name, description, dates or status_id on Cloud; add_test_cases_to_run appends cases on Cloud.
Execution results: get_test_execution returns one by execution_id, with test_run_keys narrowing the search on Data Center, and list_executions_by_cycle returns execution status, executedBy and actualEndDate per case in a cycle on Cloud.
Search: search_test_cases_by_folder takes folder_path with max_results, and search_test_runs filters on project_key and folder, with folder_id available as a direct numeric lookup that skips path resolution.
Removal on Data Center: delete_test_case and delete_test_run.

Requirements

A Zephyr Scale instance reachable at ZEPHYR_BASE_URL, a Zephyr API key as ZEPHYR_API_KEY, and Jira credentials as JIRA_USERNAME and JIRA_API_TOKEN.

Setup effort

One command plus a key — npm install -g zephyr-scale-mcp-server, then supply credentials