Docs

DocsCLI

CLI

Submit automation results from any CI with @testbeaver/cli. Point it at a JUnit XML report, or call the same endpoints as the API. Agents can still manage cases and ad-hoc results via MCP.

1. Install

npm install -D @testbeaver/cli

2. Auth

Set TESTBEAVER_API_TOKEN (cr_… from Account). Optionally set TESTBEAVER_PROJECT_ID so you can omit --project.

export TESTBEAVER_API_TOKEN=cr_…
export TESTBEAVER_PROJECT_ID=…

3. Emit JUnit XML

Configure your runner, then see the framework guides for copy-paste setup:

4. Submit results

# after your tests write results.xml
testbeaver results submit \
  --project "$TESTBEAVER_PROJECT_ID" \
  --name "CI smoke" \
  --env staging \
  --results results.xml

The CLI parses the report, matches or creates cases, starts a run, records pass/fail/untested, and closes the run. Use --no-create to only update existing cases, or --no-close to leave the run open.

How cases are matched

  1. If the test name (or a JUnit property) contains @testbeaver:<caseId>, that case is used.
  2. Otherwise the CLI uses an automation key classname.name, stored as tag auto:… on the case.
  3. If no case exists for that key, one is created (unless --no-create).

Commands

  • testbeaver results submit --project … --results <file|glob> — recommended CI path
  • testbeaver projects list
  • testbeaver cases list --project <id>
  • testbeaver runs start|close|summary
  • testbeaver results record|list — fine-grained updates without JUnit

Successful commands print JSON on stdout.

API

Prefer raw HTTP? See the API docs.