chore: release v0.8.143

[jrusso1020]

Prepare v0.8.143 from current main (2dd708e0a).

The main reason for this cut is #5290 together with #5280: distributed renders no longer extract every video frame in the plan step. Each chunk extracts only the frames it shows. On AWS Lambda the plan step for a 26-minute clip drops from about 395 s to about 20 s. The internal render backtest needs a published producer with this change.

Release notes cover all 47 pull requests merged since v0.8.142 in plain language. Both performance changes are listed under Performance. Package and plugin versions are updated together.

No visible change

This PR changes release metadata and documentation only. The features and fixes described in the notes are already on main.

Validation

• bun run release:prepare 0.8.143 on the current main commit, after fetching tags so the range starts at v0.8.142.
• Compared the release notes against every merged pull request since v0.8.142 (git log v0.8.142..2dd708e0a, 47 commits, each listed once).
• Earlier changelog entries are unchanged. The diff to docs/changelog.mdx only adds the v0.8.143 block.
• Local tag deleted. The publish workflow creates v0.8.143 at the merge SHA.

🤖 Generated with Claude Code

[mintlify]

Preview deployment for your docs. Learn more about Mintlify Previews.

Project	Status	Preview	Updated
hyperframes	🟢 Ready	View Preview	Oct 8, 2026, 11:04 PM

💡 Tip: Enable Automations to automatically generate PRs for you.

[jerrai-bot-heygen]

Exact-head release metadata review: all 20 changed files are the 13 package versions, five plugin manifests, release notes and the prepended changelog block; each version changes 0.8.142→0.8.143. I compared the 47 note PR references with the 47 commits from v0.8.142 to base 2dd708e0a and found no missing or duplicate numbers. #5280/#5290 summaries match their merged changes; the quoted 26-minute speedup is author-reported, not reviewer-reproduced. This approval covers metadata correctness, not publishing, backtesting, or deployment. Hosted tests were partly still running at review time; I did not run local tests. — Jerrai

[github-actions]

Edit accuracy: accurate 2059 (base branch 2059), smooth 1590 of those

The gate passes.
Smoothness is reported in the artifact, not gated. A case fails only if it fails 2 of 3 runs.

Quarantined, measured but not gated (0)

Unstable (1)

• crop-none-px-r30-nested-z50: tracking 0.05, pressJump 0, drop 40.14, reload 40.11, render 34.78, renderKey -, undo true, teleport true / tracking 0.05, pressJump 0, drop 0.14, reload 0.11, render 0.3, renderKey -, undo true, teleport true / tracking 0.05, pressJump 0, drop 0.14, reload 0.11, render 0.3, renderKey -, undo true, teleport true