kinora

CLI upload

Upload a Playwright json report to kinora from CI, or bulk-import a backlog, with @kinora/cli.

@kinora/cli uploads a Playwright json report to kinora. Use it when you can’t run the reporter inline (results produced in one CI job, uploaded from another), or to bulk-import a backlog of historical reports.

Upload a single report

Produce results.json with Playwright’s built-in json reporter:

// playwright.config.ts
reporter: [['json', { outputFile: 'results.json' }]]

Then upload it:

npx @kinora/cli upload results.json --project web-app --token <project-token>

Create a project API token in the dashboard under Settings → Workspace. Auth can also come from the environment (KINORA_TOKEN, KINORA_URL). Self-hosting? Point at your server with --url / KINORA_URL - see Self-hosting.

Bulk import (historical backfill)

Import every *.json report under a directory in one go, useful for seeding history from an existing archive:

npx @kinora/cli import ./reports --project web-app --token <project-token> --concurrency 8
  • Billing follows test execution date, so runs from past periods are free and don’t consume your current quota; only current-period runs count.
  • Imported runs skip alerts (no notification spam).
  • Traces are not uploaded by import: a JSON-only archive has no trace.zip next to it. History (pass rates, trends, flaky detection) still works without them.

Options

--project <slug>      required, target project slug
--token <token>       project API token (or env KINORA_TOKEN)
--url <url>           server base URL (or env KINORA_URL; default: hosted cloud, set for self-host)
--name <name>         project display name (default: slug)
--git-sha <sha>
--git-branch <branch>
--git-repo-url <url>  remote URL (https://github.com/org/repo) to link shas to commits
--ci-provider <name>
--ci-run-url <url>
--ci-run-number <n>
--git-base-branch <b> PR base branch (auto-detected in CI); powers "regression vs base"
--pr-comment          post/update a summary on the GitHub PR or GitLab MR (see PR comments guide)
--pr-label <label>    distinguish matrix legs that share one PR
--pr-policy <policy>  always (default) | on-failure (no new comment on green runs)
--upload-attachments <kinds>
                      comma-separated: trace, snapshot (both default), video, screenshot.
                      snapshot = images of a failed screenshot assertion (comparison view)
--group-key <key>     sharded report: the run its shards are assembled into
                      (default: env KINORA_GROUP_KEY, then GITHUB_RUN_ID / CI_PIPELINE_ID)
--no-group            upload a sharded report as a run of its own
--concurrency <n>     parallel uploads for bulk import (default: 6)
-h, --help

On GitHub Actions and GitLab CI, git and ci metadata (including the repo URL) auto-detect from the standard GITHUB_* / CI_* env vars; the flags override them. Pass the flags explicitly on other CI providers.

CI example (GitHub Actions)

- run: PLAYWRIGHT_JSON_OUTPUT_NAME=results.json npx playwright test --reporter=json
- run: npx @kinora/cli upload results.json --project web-app
  if: always()
  env:
    KINORA_TOKEN: ${{ secrets.KINORA_TOKEN }}
    # git + ci metadata (sha, branch, repo URL, run link) auto-detect from GITHUB_*

CI example (GitLab CI)

e2e:
  script:
    - PLAYWRIGHT_JSON_OUTPUT_NAME=results.json npx playwright test --reporter=json
  after_script:
    # KINORA_TOKEN is a masked CI/CD variable; git + ci metadata auto-detect from CI_*
    - npx @kinora/cli upload results.json --project web-app

For posting a summary on the pull request, see PR comments.

Sharded runs

Upload each shard’s report from its own job: the JSON report records which shard produced it (--shard), and kinora assembles the shards of one CI run into a single run, exactly like the reporter. Grouping, partial and incomplete runs, and re-runs work as described in Sharded runs.

- run: PLAYWRIGHT_JSON_OUTPUT_NAME=results.json npx playwright test --shard=${{ matrix.shard }} --reporter=json
- run: npx @kinora/cli upload results.json --project web-app
  if: always()
  env:
    KINORA_TOKEN: ${{ secrets.KINORA_TOKEN }}

kinora import never groups: each imported report stays a run of its own.