Skip to content

v1.24.7

Choose a tag to compare

View all tags
@github-actions github-actions released this 27 Jul 12:24
Immutable release. Only release title and notes can be modified.
v1.24.7
8472b87
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

kesha-voice-kit v1.24.7 (engine)

What's Changed

Reliability

  • fix(models): download_verified is now atomic — downloads land in a .part sibling and are renamed only after the SHA-256 check passes, so an interrupted install can no longer leave a truncated model that looks complete (#619)
  • fix(models): orphaned download staging is swept on install, reclaiming disk from earlier interrupted runs (#632)
  • fix(transcribe): the chunk-overlap tail is clamped to a character boundary, fixing corrupted text at chunk seams for multi-byte scripts (#616)

CLI

  • fix(cli): record and list_voices spawns are guarded, directories are rejected with a clear error instead of a confusing failure, and unknown subcommands now suggest the closest match (#643)

Raycast extension

  • fix(raycast): the silence clock keeps running when the audio meter stops reporting, so auto-stop still fires (#646)
  • fix(raycast): actionable onboarding — brew-first setup hint, preflight checks, and an early signal when the mic never produces samples (#641)
  • fix(raycast): cancelling transcription no longer races the transcribing toast (#615)

Docs

  • Install cost expectations are stated up front and several onboarding dead ends are fixed (#642)
  • MCP prerequisites, the PATH caveat, and the absolute-path contract for transcribe_audio are documented (#640)

No breaking changes and no public API changes. The remainder of this release is internal:
TTS/SSML and ONNX refactors, a large test-suite cleanup, and dependency bumps.

Install / upgrade

bun add -g @drakulavich/kesha-voice-kit@latest
kesha install

Verify release assets

Download the published binaries and checksums, then verify them:

gh release download v1.24.7 -p SHA256SUMS -p kesha-release-manifest.json -p '*.sigstore.json' -p 'kesha-*' -p 'say-*'
sha256sum -c SHA256SUMS

cosign verify-blob \
  --bundle kesha-engine-darwin-arm64.sigstore.json \
  --certificate-identity "https://github.com/drakulavich/kesha-voice-kit/.github/workflows/build-engine.yml@refs/tags/v1.24.7" \
  --certificate-oidc-issuer https://token.actions.githubusercontent.com \
  kesha-engine-darwin-arm64

The release also includes packaging metadata in kesha-release-manifest.json and a source SBOM in kesha-voice-kit-v1.24.7.spdx.json.

Assets 23
Loading