v1.1.0

The explainer now speaks in the voice you saved for your channel.

• voice is empty by default. Leave it out and the explainer narrates in the voice saved for the channel on bisque.cloud, else the account's. An account with no saved voice still gets kokoro:af_heart, as before. Set voice to keep a fixed one; it always wins.
• If bisque.cloud cannot be read, the run fails before anything is narrated or published, rather than publish in a voice that may not be yours. The same goes for a saved voice that cannot be spoken on a runner.
• A slide is reused only in the voice it was spoken in. Change the channel voice and the next publish narrates it again on its own. Audio published before this release records no voice, so it is narrated again once, the next time you publish that presentation, and reused as usual after that.
• Speech models are cached by what was installed, since the voice is known only once the run reads your account. A Studio voice (qwen3-*) is a multi-gigabyte download on the first run.

- uses: bisque-cloud/presenter@v1
  with:
    anthropic-api-key: ${{ secrets.ANTHROPIC_API_KEY }}
    bisque-api-key: ${{ secrets.BISQUE_API_KEY }}
    bisque-user-id: ${{ secrets.BISQUE_USER_ID }}

Full Changelog: v1.0.0...v1.1.0

v1.0.0

First stable release of the Release Explainer action. The inputs in action.yml are now a compatibility promise: v1 tracks the newest 1.x, and nothing in 1.x will rename or remove an input or change what an existing workflow does.

- uses: bisque-cloud/presenter@v1
  with:
    anthropic-api-key: ${{ secrets.ANTHROPIC_API_KEY }}
    bisque-api-key: ${{ secrets.BISQUE_API_KEY }}
    bisque-user-id: ${{ secrets.BISQUE_USER_ID }}

Same commit as v0.1.1. The 0.1.x tags stay published so existing workflows pinned to them keep resolving.

Full Changelog: v0.1.1...v1.0.0

v0.1.1

Full Changelog: v0.1.0...v0.1.1

Release Explainer v0.1.0 (pre-release)

The first tagged version of the Release Explainer action, marked pre-release because it has not yet completed a run on a hosted runner.

on:
  release:
    types: [published]
permissions:
  contents: write
jobs:
  explain:
    runs-on: ubuntu-latest
    steps:
      - uses: bisque-cloud/presenter@v0.1.0
        with:
          model: anthropic/claude-sonnet-4-6
          api-key: ${{ secrets.ANTHROPIC_API_KEY }}
          bisque-api-key: ${{ secrets.BISQUE_API_KEY }}
          bisque-user-id: ${{ secrets.BISQUE_USER_ID }}

model is any provider/model id from models.dev. OpenCode runs it, so switching models is one line.

What is verified. Gathering a release, selecting the diff, the key proxy, the skill hash, and the presentation structure check each ran against real data. One full authoring run produced a 7-slide presentation with downloaded fonts that passes the format contract.

What is not. Narration and publishing have never run on a GitHub runner, and nothing has run on ubuntu-latest. That is what this pre-release is for.

Known limits. The action installs bisque-voice with a shell installer from download.bisque.today; a pinned, checksummed release asset replaces that before v1.0.0.