Skip to content

Releases: vizzly-testing/cli

✨ v0.30.0

02 May 06:03

Choose a tag to compare

Vizzly CLI v0.30.0 Release Notes

What's Changed

Added

  • CLI Visual Context Workflows: New read-only vizzly context commands for programmatic access to build, comparison, screenshot, similar, and review queue data. Supports both cloud context and local workspace context from .vizzly/, enabling integration with CI/CD pipelines and custom workflows.
  • JSON Output Reference: Complete documentation of JSON output format for all CLI operations in docs/json-output.md.

Changed

  • Storybook 10 Support: Updated and verified full compatibility with Storybook 10.3.6. The bundled example app now uses Storybook 10, and local TDD metadata is properly sanitized for iframe environments.
  • TypeScript 6.0: Upgraded TypeScript from 5.9.3 to 6.0.3.
  • Build Tool Alignment: Updated Vite to v8 and plugin-react to v6 for consistency across the reporter UI toolchain.

Fixed

  • Stale TDD Image Loads: Fixed an issue where baseline, current, and diff images weren't refreshing properly in the reporter UI. Added cache-busting via query parameters and no-cache response headers to ensure fresh image data.
  • Storybook Story URL Capture: Fixed Storybook screenshot metadata to use proper story paths (?path=/story/...) instead of iframe URLs, ensuring story links render correctly in the Vizzly dashboard.
  • Storybook CI Preview Uploads: Added preview upload step to Storybook SDK E2E workflow in CI.

Full Changelog: v0.29.6...v0.30.0

✨ v0.29.6

17 Feb 13:46

Choose a tag to compare

Vizzly CLI v0.29.6

What's Changed

Added

  • Full-page screenshot capture is now the default for Static Site and Storybook SDKs. Teams can opt out with --no-full-page or by setting screenshot.fullPage: false in config if they prefer viewport-only captures.

Fixed

  • Fixed build auto-detection for plugin-created builds (e.g., vizzly storybook). TestRunner now writes session context (buildId, branch, commit, parallelId) for non-TDD builds, enabling vizzly preview to properly locate and process them.

Breaking Changes

Default screenshot behavior changed to full-page capture in Static Site and Storybook SDKs. If you want to keep capturing viewport-only screenshots, use:

  • Config: screenshot: { fullPage: false }
  • CLI: --no-full-page

Full Changelog: v0.29.5...v0.29.6

✨ v0.29.5

16 Feb 23:09

Choose a tag to compare

Vizzly CLI v0.29.5

What's Changed

Added

  • Added explicit CLI opt-out flag --no-full-page for controlling screenshot capture behavior in Static Site and Storybook SDKs

Changed

  • ⚠️ Breaking Change: Static Site and Storybook SDKs now default to full-page screenshots instead of viewport-only. This captures the entire scrollable content by default, which better matches user expectations for static pages and Storybook stories
    • To restore the previous viewport-only behavior, use screenshot: { fullPage: false } in your config or the --no-full-page CLI flag

Fixed

  • Fixed vizzly preview command failing to auto-detect builds created by plugin commands (like vizzly storybook). The session breadcrumb is now properly written for non-TDD builds, ensuring build context is preserved

Full Changelog: storybook/v0.7.0...v0.29.5

✨ v0.29.4

16 Feb 19:56

Choose a tag to compare

Vizzly CLI v0.29.4

What's Changed

Added

  • Page URL capture in all JS client SDKs: Screenshots now automatically capture the current page URL as a url property. This enables backend integration with deploy preview pages. The URL is captured in Static Site, Storybook, Vitest, and Ember SDKs and can be overridden by users if needed.

Fixed

  • Fixed double-nested properties in client screenshot payload: The client SDK was incorrectly double-nesting the properties object when SDKs passed { properties: { url } }. The payload is now properly flattened so properties reach the server correctly. Also prevents SDK-internal options (fullPage, threshold) from leaking into the properties object.
  • Fixed local TDD dimension mismatch handling: Local TDD mode was silently auto-overwriting baselines when honeydiff detected a width mismatch. Now dimension mismatches are treated as failed comparisons, allowing teams to review and explicitly accept/reject the changes in the local TDD UI instead of silently overwriting baselines.
  • Updated 7 dependencies: Routine dependency updates including dotenv, glob, Biome, TanStack React Query, TypeScript types, Vite React plugin, and rimraf.

Full Changelog: v0.29.3...v0.29.4

📚 Storybook Plugin v0.7.0

16 Feb 20:00

Choose a tag to compare

[0.7.0] - 2026-02-16

What's Changed

Added

  • Page URL is now automatically captured with each screenshot and sent as a property. This enables the Vizzly dashboard to link screenshots to their corresponding Storybook story URLs for easier navigation.

Full Changelog: storybook/v0.6.0...storybook/v0.7.0

✨ v0.29.3

14 Feb 04:52

Choose a tag to compare

Vizzly CLI v0.29.3

What's Changed

Changed

  • Optimized TDD dashboard real-time updates — Reduced performance overhead in the interactive dashboard by consolidating Server-Sent Events (SSE) connections. Previously, multiple components each created their own EventSource connection. Now the dashboard uses a singleton SSE connection managed by a React context provider, significantly improving performance on large test suites.
  • Incremental event payloads — The TDD dashboard now sends only changed comparison data instead of re-sending the entire report on every update. For a 139-screenshot suite, this reduces per-event payload from 50-200KB down to 0.5-2KB, resulting in faster updates and less bandwidth consumption.

Full Changelog: storybook/v0.6.0...v0.29.3

📚 Storybook Plugin v0.6.0

13 Feb 00:24

Choose a tag to compare

[0.6.0] - 2026-02-13

What's Changed

No changes to Storybook plugin in this release.

Full Changelog: storybook/v0.5.0...storybook/v0.6.0

✨ v0.29.2

11 Feb 18:35

Choose a tag to compare

v0.29.2

What's Changed

Added

  • --org flag for disambiguating projects across organizations: The builds and comparisons commands now accept an --org <slug> option to filter results by organization. This is useful when multiple projects in different organizations share the same project slug.
    vizzly builds --project storybook --org my-org
    vizzly comparisons --name "Button" --project storybook --org my-org

Fixed

  • TDD SSE performance: Fixed a critical performance issue where TDD mode was broadcasting unnecessarily large Server-Sent Events (SSE) updates. Heavy comparison data (diffClusters, intensityStats, etc.) is now split into a separate file and loaded on-demand when the fullscreen viewer opens. This reduces per-update bandwidth by 79% (from 735 KB to 153 KB per update) and cumulative broadcast traffic from ~97 MB to ~20 MB for typical test runs. (#214)

Full Changelog: v0.29.1...v0.29.2

✨ v0.29.1

10 Feb 02:53

Choose a tag to compare

Vizzly CLI v0.29.1

What's Changed

Added

  • Honeydiff analysis data — Expose detailed visual analysis metrics (ssim_score, gmsd_score, cluster_metadata, fingerprint_hash, diff_regions) in comparison output for both JSON and human-readable formats
  • Screenshot URLs in comparisons — Include baseline, current, and diff image URLs when viewing comparison details
  • Comparison counts in build list — Display summary of comparisons (3 changed · 2 new · 25 identical) in the builds list output
  • Approval badges in build list — Show colored approval status badges (APPROVED, PENDING, REJECTED) next to completed builds
  • Diff percentage display — Show percentage difference for each comparison in the builds output
  • Classification labels — Display honeydiff classification in verbose mode to help identify types of visual changes

Fixed

  • Undefined comparison names — Fixed issue where comparison names could appear as "undefined" by using fallback chain: name → current_name → id
  • JSON compatibility — Improved robustness with older API responses by adding current_name fallback for the name field

Full Changelog: v0.29.0...v0.29.1

✨ v0.29.0

09 Feb 06:49

Choose a tag to compare

v0.29.0

What's Changed

Added

  • Honeydiff analysis data in comparisons output: The comparisons and builds commands now surface detailed visual analysis metrics from honeydiff:

    • SSIM and GMSD scores for quantified visual differences
    • Cluster classification and metadata for intelligent difference categorization
    • Fingerprint hash for change fingerprinting
    • Diff regions data for precise difference localization
    • Available in both --json output and --verbose displays
  • Image URLs in comparison details: All comparison outputs now include resolved URLs for baseline, current, and diff images, handling various API response formats consistently

  • Enhanced verbose output: Single comparison display and list views now show classification labels and detailed analysis metrics for better visual debugging

Full Changelog: v0.28.1...v0.29.0