Base URL: https://api.elevenlabs.io/v1

POST /text-to-speech/{voice_id}           # Generate audio
POST /text-to-speech/{voice_id}/stream    # Stream audio
GET  /voices                              # List voices
GET  /voices/{voice_id}                   # Voice details
GET  /user                                # Usage/quota
POST /speech-to-speech/{voice_id}         # Voice conversion

import { ElevenLabsClient } from 'elevenlabs';

const client = new ElevenLabsClient({
  apiKey: process.env.ELEVENLABS_API_KEY
});

async function generateNarration(
  text: string,
  voiceId: string = 'Rachel'
): Promise<Buffer> {
  const audio = await client.generate({
    voice: voiceId,
    text: text,
    model_id: 'eleven_multilingual_v2',
    voice_settings: {
      stability: 0.5,
      similarity_boost: 0.8,
      style: 0.0,
      use_speaker_boost: true
    }
  });

  // Convert stream to buffer
  const chunks: Buffer[] = [];
  for await (const chunk of audio) {
    chunks.push(chunk);
  }
  return Buffer.concat(chunks);
}

interface VoiceSettings {
  stability: number;        // 0.0-1.0 (lower = more expressive)
  similarity_boost: number; // 0.0-1.0 (higher = closer to original)
  style: number;           // 0.0-1.0 (v2 models only)
  use_speaker_boost: boolean; // Clarity enhancement
}

// Recommended settings by content type
const VOICE_PRESETS = {
  narration: { stability: 0.65, similarity_boost: 0.8, style: 0.0 },
  conversational: { stability: 0.4, similarity_boost: 0.75, style: 0.2 },
  dramatic: { stability: 0.3, similarity_boost: 0.9, style: 0.5 },
  professional: { stability: 0.8, similarity_boost: 0.85, style: 0.0 },
  energetic: { stability: 0.35, similarity_boost: 0.85, style: 0.4 }
};

function framesToMs(frames: number, fps: number = 30): number {
  return Math.round((frames / fps) * 1000);
}

function msToFrames(ms: number, fps: number = 30): number {
  return Math.round((ms / 1000) * fps);
}

// Examples
framesToMs(90, 30);   // 3000ms (3 seconds at 30fps)
framesToMs(150, 30);  // 5000ms (5 seconds at 30fps)
msToFrames(2500, 30); // 75 frames

Speaking Speed       WPM     Words/30s    Use Case
----------------------------------------------------------
Slow (dramatic)      100     50           Hooks, reveals
Normal narration     130-150 65-75        Standard content
Conversational       150-170 75-85        Tutorials, demos
Fast (excited)       170-190 85-95        Features, energy
Very fast            200+    100+         Avoid (unclear)

import { Audio, Sequence, useVideoConfig } from 'remotion';

interface NarrationProps {
  audioUrl: string;
  startFrame: number;
  volume?: number;
}

export const Narration: React.FC<NarrationProps> = ({
  audioUrl,
  startFrame,
  volume = 1
}) => {
  return (
    <Audio
      src={audioUrl}
      startFrom={0}
      volume={volume}
    />
  );
};

// Usage in a scene
export const NarratedScene: React.FC = () => {
  return (
    <>
      <Sequence from={0} durationInFrames={150}>
        <HookScene />
        <Narration audioUrl="/audio/hook-narration.mp3" startFrame={0} />
      </Sequence>

      <Sequence from={150} durationInFrames={300}>
        <DemoScene />
        <Narration audioUrl="/audio/demo-narration.mp3" startFrame={150} />
      </Sequence>
    </>
  );
};

function estimateCost(
  text: string,
  model: 'multilingual' | 'turbo' | 'flash' = 'multilingual'
): number {
  const chars = text.length;
  const costPer1K = {
    multilingual: 0.30,
    turbo: 0.18,
    flash: 0.08
  };
  return (chars / 1000) * costPer1K[model];
}

# Required
ELEVENLABS_API_KEY=xi_xxxxxxxxxxxxxxxxxxxx

# Optional
ELEVENLABS_MODEL_ID=eleven_multilingual_v2
ELEVENLABS_DEFAULT_VOICE=21m00Tcm4TlvDq8ikWAM