> ## Documentation Index
> Fetch the complete documentation index at: https://docs.varg.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Sync Lipsync

> High-quality lip synchronization — apply speech audio to existing video

## Overview

Sync (by Synchronize Labs) applies lip synchronization to existing videos. Provide a video and an audio file, and the model will make the person in the video appear to speak the audio. Two quality tiers are available.

| Model ID      | Quality | Speed    | Credits | \~Cost |
| ------------- | ------- | -------- | ------- | ------ |
| `sync-v2-pro` | Best    | \~60-90s | 80      | \$0.80 |
| `sync-v2`     | Good    | \~45-60s | 50      | \$0.50 |
| `lipsync`     | Basic   | varies   | 50      | \$0.50 |

## Quick start

<CodeGroup>
  ```typescript SDK theme={null}
  import { createVarg } from "vargai/ai"

  const varg = createVarg({ apiKey: process.env.VARG_API_KEY! })

  const result = await varg.videoModel("sync-v2-pro").generate({
    videoUrl: "https://example.com/talking-head.mp4",
    audioUrl: "https://example.com/speech.mp3",
  })

  console.log(result.video.url)
  ```

  ```bash cURL theme={null}
  curl -X POST https://api.varg.ai/v1/video \
    -H "Authorization: Bearer $VARG_API_KEY" \
    -H "Content-Type: application/json" \
    -d '{
      "model": "sync-v2-pro",
      "prompt": "",
      "files": [
        { "url": "https://example.com/talking-head.mp4" },
        { "url": "https://example.com/speech.mp3" }
      ]
    }'
  ```
</CodeGroup>

## Parameters

<ResponseField name="files" type="array" required>
  Two files: one video and one audio. The gateway auto-detects file types by extension.
</ResponseField>

<Note>
  Sync models don't use prompt, duration, or aspect\_ratio parameters. The output matches the input video dimensions and the audio duration.
</Note>

## Full talking head pipeline

The typical workflow: generate character image, create video, generate speech, apply lipsync.

```tsx theme={null}
// 1. Generate character
const character = Image({
  model: varg.imageModel("soul"),
  prompt: "professional presenter, neutral expression",
  aspectRatio: "9:16",
})

// 2. Generate base video
const baseVideo = Video({
  model: varg.videoModel("kling-v3"),
  prompt: { text: "subtle head movement, blinking", images: [character] },
  duration: 10,
})

// 3. Generate speech
const speech = Speech({ model: varg.speechModel("eleven_v3"), text: "Welcome to our product demo..." })

// 4. Apply lipsync
const talkingHead = Video({
  model: varg.videoModel("sync-v2-pro"),
  prompt: { video: baseVideo, audio: speech },
})
```

## Pricing

| Model         | Credits | USD    |
| ------------- | ------- | ------ |
| `sync-v2-pro` | 80      | \$0.80 |
| `sync-v2`     | 50      | \$0.50 |
| `lipsync`     | 50      | \$0.50 |

## Tips

* **Pro is recommended** for production content. The quality difference is noticeable, especially around mouth movements.
* **Input video should have a clear face** — front-facing, well-lit, with the face occupying a good portion of the frame.
* **Audio quality matters** — clean speech audio produces much better lipsync results.
* **Combine with ElevenLabs** for the full pipeline: TTS -> Sync V2 Pro.

## Related models

<CardGroup cols={2}>
  <Card title="VEED Fabric" icon="comment" href="/models/lipsync/veed-fabric">
    Simpler pipeline — image + audio, no video needed.
  </Card>

  <Card title="OmniHuman" icon="person" href="/models/lipsync/omnihuman">
    Full-body animation, not just lips.
  </Card>

  <Card title="ElevenLabs" icon="microphone" href="/models/speech/elevenlabs">
    Generate speech audio for lipsync.
  </Card>
</CardGroup>