@lincyaw/slidev-addon-speech

Speech services: speech synthesis and transcription through the operator's engines, the synthesis cache, and presmium-speech.

Part of Presmium, which turns lectures written in MyST Markdown into web pages, slides, Jupyter notebooks and narrated lectures.

Install

pnpm add -D @lincyaw/slidev-addon-speech

List it in the headmatter of the documents that use it:

plugins:
  - '@lincyaw/slidev-addon-speech'

It is a Presmium plugin and a Slidev addon: its components work on the web page and in the slides. It needs @lincyaw/cli and the renderers in the same project, and declares the Presmium packages it shares with them as peer dependencies (Presmium 1.x).

What it provides

• The SPEECH service in Node and in the browser, with the cache .presmium/speech/ (commit it, and others build without credentials).
• Speech is marked as AI generated in the player and in the audio's metadata; cloned voices need a recorded consent (presmium-speech consent).

Configuration

The operator defines engines in environment variables: PRESMIUM_SPEECH_<ID>_URL, _PROVIDER and, optionally, _KEY. Documents name the engine's id only, for example presmium: { presmium.speech: { engine: volcengine } }.

Documentation

• Author guide (Chinese: site/guide/lecture.md)
• Directive reference (Chinese): every directive with its options and an example

License

Apache-2.0