Skip to content

chore: version packages#69

Merged
LEVI-RIVKIN merged 1 commit into
mainfrom
changeset-release/main
Jul 8, 2026
Merged

chore: version packages#69
LEVI-RIVKIN merged 1 commit into
mainfrom
changeset-release/main

Conversation

@github-actions

@github-actions github-actions Bot commented Jul 8, 2026

Copy link
Copy Markdown
Contributor

This PR was opened by the Changesets release GitHub action. When you're ready to do a release, you can merge this and the packages will be published to npm automatically. If you're not ready to do a release yet, that's fine, whenever you add more changesets to main, this PR will be updated.

Releases

@perspective-ai/sdk@1.14.0

Minor Changes

  • fe18680: Replace the loading skeleton with the interview's real chrome.

    Instead of a skeleton that mirrors the interview UI's internal layout (welcome
    card, message, input pill) — which drifts whenever that UI changes — the
    loading state now renders what the interview actually looks like from the
    outside in:

    • the research's scene image (fetched from
      {host}/interview/{researchId}/scene-image) behind
    • a frosted translucent card (55% panel + backdrop blur, matching the
      app's card treatment), with
    • a circular loader centered in the card (40px) — a conic-gradient arc
      with a fading tail over a faint track, tinted with the brand primary
      resolved with the SDK's usual precedence: local brand.primary override →
      API embed config (primaryColor / darkPrimaryColor) → Perspective
      default (#7c3aed / #a78bfa). As essential loading motion it is not
      disabled by prefers-reduced-motion.

    The scene image is applied only after it has actually loaded, layered over
    the app's default surface color (#f5f2f0 light, #15171e dark, or a custom
    brand.bg) — so a research without a scene (404), a slow network, or a
    blocked request cleanly degrades to a solid color.

    Because the loading state renders in the host page DOM (an overlay sibling of
    the cross-origin iframe), responsiveness is driven by CSS container
    queries
    on the slot the consumer provides — not viewport media queries. The
    card's sizing mirrors the interview app's own container queries (full-bleed
    below 672px width; a centered max-width: 672px box above it, with vertical
    padding stepping 4px → 48px → 80px at the app's 448px/768px container-height
    breakpoints), so the card doesn't jump at handoff. Breakpoints are in px, not
    the app's rem: the overlay lives in the host page, so rem would resolve against
    the host's (uncontrolled) root font-size, whereas the card must match the
    iframe's fixed pixels.

    LoadingOptions gains optional researchId and host (used for the scene
    URL) and apiConfig (used for the loader tint); all embed types pass them
    automatically.

    A new prefetchSceneImage(researchId, host?) export warms the scene image on
    intent signals — the SDK wires it to float-bubble hover/focus and to
    data-perspective-popup/-slider trigger elements — so deferred embeds open
    with the scene already cached.

@perspective-ai/sdk-react@1.14.0

Patch Changes

  • Updated dependencies [fe18680]
    • @perspective-ai/sdk@1.14.0

@LEVI-RIVKIN LEVI-RIVKIN merged commit 49c3e70 into main Jul 8, 2026
5 checks passed
@LEVI-RIVKIN LEVI-RIVKIN deleted the changeset-release/main branch July 8, 2026 19:50
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant