speakasm v0.2.1

Changelog

• 5e4b893 Add Chrome and Firefox extensions with site-hosted CRX/XPI builds.
• 5df0303 Add extension update checks and configurable site/bridge settings.
• 987573f Bump version to 0.2.0.
• ecf7c11 Bump version to 0.2.1.
• 41f8cc8 Copy vendor package-lock into the Docker build for npm ci.
• 6531e11 Fix CI: CRX sandbox, Trivy signing key skip, and native-arch Docker vendor stage.
• b1801ea Fix badge orphan-repo identity and pin speakasm npm ci.
• 6546fc5 Fix workflow YAML broken by automated hardening edits.
• ff422ef Harden CI for OpenSSF and fix flaky Pages/Badges.
• f084d03 Harden HTML/DOCX text extraction against CodeQL sanitizer findings.
• f3bf8c5 Harden repository against GitHub Actions supply-chain attacks.
• 266d92c Install Syft in Release so GoReleaser can emit CycloneDX SBOMs.
• fba995b Override sharp to 0.35.4 to clear Trivy HIGH finding.
• d251c35 Publish CycloneDX SBOMs and cosign signatures atomically for immutable releases.
• 6147865 Restore Harden Runner as the first step in workflow jobs.
• dff5d76 Run CI tests and server smoke on Linux, macOS, and Windows.
• 5aaa315 Track vendor package-lock so Pages can npm ci.

Docker

docker pull ghcr.io/quad4-software/speakasm:v0.2.1
docker run --rm -p 8080:8080 ghcr.io/quad4-software/speakasm:v0.2.1

Multi-arch images: linux/amd64, linux/arm64. Image includes Kokoro ONNX + vendor WASM.

Binary

Unpack the archive for your OS/arch, then point -web at a web/ tree with models
(make assets after cloning), or use the container image for a full offline stack.

Extensions

Browser builds are attached to this release and also published at
https://speakasm.quad4.io/build/

• speakasm-chrome.crx / speakasm-chrome.zip
• speakasm-firefox.xpi

speakasm v0.1.0

Changelog

• d49c481 Add EditorConfig for consistent formatting across editors.
• 128030c Add GoReleaser and GHCR publishing for tagged releases.
• df921dd Add Lighthouse CI with a shell-only asset fetch path.
• 928ee75 Add OpenSSF Scorecard and harden GitHub Actions supply chain.
• a617cb0 Add Playwright screenshot capture for docs and CI artifacts.
• cd3a1cd Add planTotal function to Kokoro runtime for sentence unit counting
• 38eb927 Add themed OpenSSF Scorecard endpoint badges.
• e6b69a1 Add themed shields.io endpoint badges and a refresh workflow.
• 92754e2 Add transparent PNG icons for each Kokoro voice.
• 4d47fde Bump checkout to 7.0.1 and configure-pages to 6.0.0.
• 30dee29 Bump github/codeql-action/upload-sarif from 4.37.4 to 4.37.9 (#3)
• f0e5787 Bump the service worker cache for the refreshed UI shell.
• f5c87d6 Clear generated PCM chunks after concat to free memory sooner.
• 1898183 Disable credential persistence on the Pages checkout step.
• 7add639 Document Docker pulls, release binaries, badges, and screenshots.
• c076cf4 Fix gocritic httpNoBody lint in server tests.
• 3a130b4 Fix local Kokoro loading, voices fallback, fonts, and viewport layout.
• 66b6db5 Guard the wave draw loop so RAF stops when the tab is already hidden.
• 943a98a Harden PWA updates with stamped shell caches and idle reload.
• 6091097 Harden the Docker image and make compose image tags configurable.
• ad17990 Initial release of speakasm: offline browser TTS via Kokoro ONNX.
• 8800bb3 Keep the UI responsive during large pastes and generation.
• cecf025 Pause the waveform animation loop while the tab is hidden.
• 7deed85 Polish mobile layout with safe areas, voice rail fades, and install tips.
• 6fc98ed Publish container images to GHCR on every master push.
• 6a05704 Remove the Install and iOS tip chrome from the UI.
• bc65e01 Replace brand assets with unique speakasm logo and OG art.
• 43a4192 Replace cleaned-text panel with a dedicated playback section.
• 429f844 Replace the voice dropdown with an icon picker and polish controls.
• 7c1f934 Reuse the replay AudioBuffer and release sources to cut playback memory churn.
• 0e4d068 Speed up Kokoro by pairing WebGPU with fp32 and WASM with q8.
• f2c62cd Sync Scorecard pin comment to codeql-action v4.37.9.
• 96bc964 Tighten the mobile shell with sticky actions and clearer playback chrome.
• 8cb58c4 Update Kokoro runtime with WebGPU audio quality checks and fallback mechanisms.

Docker

docker pull ghcr.io/quad4-software/speakasm:v0.1.0
docker run --rm -p 8080:8080 ghcr.io/quad4-software/speakasm:v0.1.0

Multi-arch images: linux/amd64, linux/arm64. Image includes Kokoro ONNX + vendor WASM.

Binary

Unpack the archive for your OS/arch, then point -web at a web/ tree with models
(make assets after cloning), or use the container image for a full offline stack.