@bjelser/tts-elevenlabs

ElevenLabs TTS provider for @bjelser/kit. Synthesizes per-beat narration with ElevenLabs voices and exposes word-level alignment for karaoke-style passage / read-along scenes.

Install

npm install @bjelser/tts-elevenlabs @elevenlabs/elevenlabs-js
# or
bun add @bjelser/tts-elevenlabs @elevenlabs/elevenlabs-js

@elevenlabs/elevenlabs-js is a peer dependency — bring your own SDK version.

Use

// docent.config.ts
import {corePlugins} from '@bjelser/core';
import {elevenlabsTtsProvider} from '@bjelser/tts-elevenlabs';

export default {
  plugins: [...corePlugins, elevenlabsTtsProvider],
};

// films/my-film.json
{
  "meta": {
    "voice": {
      "provider": "elevenlabs",
      "voice": "rachel",
      "model": "eleven_multilingual_v2"
    }
  }
}

Set ELEVENLABS_API_KEY in the environment.

Capabilities

• nativeAlignment: 'word' — emits per-word timestamps the renderer can bind to highlight / karaoke effects
• voiceCloning: true (via ElevenLabs' Instant Voice Cloning)
• languages: per ElevenLabs' published list

A scene plugin that declares requiresTtsCapabilities: { nativeAlignment: 'word' } will pass the cross-bind check when this provider is registered.

License

MIT